Making Sense of Payments Policy in the InformationAge

ESSAYS

Making Sense of Payments Policyin the Information Age

RONALD J. MANN* **

TABLE OF CONTENTS

INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 634

I. PAYMENTS POLICY IN GENERAL . . . . . . . . . . . . . . . . . . . . . . . . . . . 636

II. THE EXISTING FRAMEWORK . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 642

A. REVERSIBILITY IN EXISTING PAYMENT SYSTEMS . . . . . . . . . . . . . 642

1. Reversibility with Checks . . . . . . . . . . . . . . . . . . . . . . . . 643

2. Reversibility with Credit Cards . . . . . . . . . . . . . . . . . . . . 647

3. Reversibility with Debit Cards . . . . . . . . . . . . . . . . . . . . . 649

B. THE EXISTING POLICY JUSTIFICATION: LIMITING IMPRUDENT

BORROWING . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 650

III. A BETTER FRAMEWORK: REDRESSING AN IMBALANCE OF LEVERAGE . . 652

A. THE ILLUSORY CREDIT/DEBIT DISTINCTION . . . . . . . . . . . . . . . . . 653

B. THE LIMITED USE OF CREDIT WITH CREDIT CARDS . . . . . . . . . . . . 656

C. LEVERAGE AND DISPUTE RESOLUTION . . . . . . . . . . . . . . . . . . . . 658

1. What Policies Does TILA Serve? . . . . . . . . . . . . . . . . . . 659

2. Is TILA Beneficial? . . . . . . . . . . . . . . . . . . . . . . . . . . . . 660

IV. APPLICATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665

A. FACE-TO-FACE TRANSACTIONS . . . . . . . . . . . . . . . . . . . . . . . . 665

* Ben H. & Kitty King Powell Chair in Law, Co-Director, Center for Law, Business & Economics,The University of Texas School of Law. I thank Kyle Logue for inspiration and Allison Mann forunceasing aid of all kinds. I received useful comments on earlier versions from Dan Keating, BobRasmussen, Jim Rogers, Norman Silber, and Jim White. I also am grateful for comments I received onpresentations of early versions of this Essay to the Institute for Monetary and Economic Studies of theBank of Japan, a Symposium on Electronic Payments at the University of Michigan School ofInformation, a seminar on commercial law at the University of Virginia School of Law, and participantsin a colloquium at the Federal Reserve Bank of New York. For convenience, all citations to provisionsof the Uniform Commercial Code without indication of date refer to the current version promulgated bythe ALI and NCCUSL.

** Copyright © 2005 Ronald J. Mann.

633

B. INTERNET TRANSACTIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . 670

V. CONCLUSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 673

INTRODUCTION

Although I had been mulling over the ideas in this Essay for quite some time,I finally was driven to put the ideas on paper by a call from a colleague oneFriday afternoon. He recently had purchased something on the Internet. Regretta-bly, the Internet merchant had never shipped the goods; apparently the merchanthad failed. My colleague had given the merchant the number from his Visa cardto pay for the transaction. Being well educated, my colleague assumed that hecould have the charge removed from his credit card statement.

When he called the toll-free service line for the bank that issued his card,however, he was surprised to hear that he could not rescind the charge becauseit had been processed as a debit card transaction, rather than a credit cardtransaction. He called me, hoping that I would tell him his bank was incorrect.Unfortunately, because the Visa card was a debit card rather than a credit card, Iwas forced to tell him no, that his bank was acting within its rights.

That anecdote crystallizes something that is deeply wrong with the currentframework of our payments policy. The portion of that policy reflected in theTruth in Lending Act (TILA) grants consumers a variety of generous rights incredit card transactions.1 By contrast, the Electronic Fund Transfer Act (EFTA),which governs debit card transactions, is relatively chary in its protections.2

Today, however, with the debit card market increasingly dominated by thePIN-less debit cards marketed by Visa and MasterCard,3 the distinction betweenthe credit card and the debit card is almost invisible to all but the mostsophisticated consumers.

This Essay explores that problem. The cause of the problem is easy to see:technology has altered the landscape of private payment institutions, and Con-gress has not updated the statutes that regulate those transactions. The solution,however, seems sufficiently difficult to warrant detailed consideration. The firstpart of the Essay starts at a high level of generality, analyzing the generalquestion of what types of considerations should inform a sophisticated pay-ments policy. There has been some substantial prior thinking on the subject,most obviously in the Uniform New Payments Code (UNPC)4 and by writers

1. For a general summary of those rights, see RONALD J. MANN, PAYMENT SYSTEMS AND OTHER

FINANCIAL TRANSACTIONS 114–18, 121–38 (2d ed. 2003). For a more detailed analysis, see infra Part II.2. For a general summary of those rights, see MANN, supra note 1, at 146–54. For a more detailed

analysis, see infra Part II.3. For an introduction to the PIN-based and PIN-less debit cards, see MANN, supra note 1, at 144–46.

For a more detailed analysis, see infra Part II.4. Because the UNPC was not adopted in any jurisdiction or finally approved by the American Law

Institute (ALI) and the National Conference of Commissioners on Uniform State Laws (NCCUSL),there is no official final version. For this project, I have relied on copies kindly loaned to me by the

634 [Vol. 93:633THE GEORGETOWN LAW JOURNAL

like Peter Alces who criticized that project.5 The central point of Part I is thatprevious analysis has failed to recognize the importance of the underlyingtransactions in which payments are made to issues ordinarily treated in the legalrules that regulate payment systems. Generally, I argue that issues of paymentspolicy need to be separated into two categories: those for which determinationof the appropriate rule is heavily influenced by the technology of the paymentssystem; and those for which determination of the appropriate rule depends forthe most part on the nature of the underlying transaction. Among other things,my argument suggests that issues like the one raised above—the ability of apurchaser to reverse a completed payment—should be driven more by transac-tional considerations. For example, I argue here that the differences in relativeleverage between merchants and consumers support a broader role for reversibil-ity of payment in consumer transactions, while issues about unauthorizedtransactions should be driven more by the nature of the technology.

To illustrate that framework in application, the remainder of the Essayconsiders the two most rapidly growing payment systems in our economy,credit and debit cards. Part II starts with a summary of the existing positive lawon the subject and a description of the normative policy that law seems toreflect: a concern about cognitive limitations that afflict consumers in debttransactions. Part III challenges the value of that framework by arguing that thenormative underpinnings of the current rules have become outmoded in severalways. Most importantly, it seems likely that many if not most credit cardtransactions no longer reflect the kinds of long-term borrowing transactions thatjustify the cognitive concerns underlying the existing policy. Similarly, thepractical distinctions between credit and debit cards have eroded to the pointwhere it is difficult to rely on that distinction as conveying anything of apparentimport to the typical consumer.6 That is not to say, of course, that the use ofcredit cards for borrowing is not an important policy issue.7 It is to say,however, that it is not sufficiently dominant in the decision to use a credit cardto exclude other relevant concerns.

That argument leads to at least two possible responses. The simplest course

Harvard Law Library. References that do not identify a particular draft are to the final draft that wasformally circulated. UNIFORM NEW PAYMENTS CODE (Permanent Editorial Bd. for the Uniform Commer-cial Code, Draft No. 3, June 2, 1983) (available at Harvard Law Library). For a detailed explanation ofthe project, see HAL S. SCOTT, NEW PAYMENT SYSTEMS: A REPORT TO THE 3-4-8 COMMITTEE OF THE

PERMANENT EDITORIAL BOARD OF THE UNIFORM COMMERCIAL CODE (1978) (available at Harvard LawLibrary).

5. See Peter A. Alces, A Jurisprudential Perspective for the True Codification of Payments Law, 53FORDHAM L. REV. 83 (1984).

6. Although I do discuss in Part IV some of the difficulties of relying on disclosures in this context,it is not the project of this Essay to consider in any general way how to construct and implement legaldistinctions in a way that will be comprehensible to consumers. I have argued elsewhere that some ofthe cognitive problems related to credit cards can be remedied by real-time disclosures to consumers.See Ronald J. Mann, Global Credit Card Use and Debt: Policy Issues and Regulatory Responses (2005)(unpublished manuscript). As I explain below, that approach seems less effective in this context.

7. I discuss that topic comprehensively in Mann, supra note 6.

2005] 635PAYMENTS POLICY

would be to repeal the TILA protections as outdated. Part III, however, buildson the framework articulated in the first part of the Essay to argue that thoserules have come fortuitously to serve other important policies,8 specifically thetransaction-related sales policies that are important in issues of reversibility. Themost obvious one is to redress the imbalance in enforcement capabilitiesbetween the typical merchant seller and the typical consumer buyer. If theexisting rules shift the burden of instituting and pursuing litigation from theconsumer to the merchant, they might enhance the likelihood that disputes inretail transactions are resolved appropriately. If so, then that burden-shiftingjustifies the TILA rules.

The problem with such a justification, at least as an explanation of existinglaw, is that the enforcement-imbalance concern applies to any payment systemthat the consumer uses. The major point cutting against broad application of thatrule is the need to maintain payment options with different characteristics, sothat merchants and their customers can choose between payment systems thatare final and those that are not final. Part IV of the Essay considers theimplications of that point in two contexts, conventional face-to-face retailpayments and Internet payments. In both contexts, I conclude that credit anddebit cards should have similar limitations on reversibility, leaving open thepossibility that developing formats like prepaid cards ultimately might be lefton the cash-equivalent side of the line. To be sure, that does limit the availabil-ity under current technology of any payment option in the Internet context thatis both practical and wholly final. In the end, however, the distinctions betweencredit and debit cards do not seem sufficient or sufficiently clear to the user tojustify a major distinction in how they are treated legally.

I. PAYMENTS POLICY IN GENERAL

Two events at the close of the twentieth century have underscored the need tothink more clearly about payments policy. The first is the proliferation ofmarkets in which credit and debit cards are used. What once was a nicheproduct designed for the payment of expenses by business travelers has nowcome into widespread use in a variety of contexts that raise differing policyconcerns. The second, related to the first, is the substantial shift in the locus ofretail payment transactions from retail, face-to-face payments in brick-and-mortar stores to remote payments for Internet purchases. Collectively, thosechanges have destabilized the system for which existing payments rules weredesigned.

As the introduction suggests, I would be remiss to discuss harmonization ofrules for payment cards without some general consideration of the propriety of

8. The pattern is a familiar one, especially in the context of common-law rules. See, e.g., OLIVER W.HOLMES, THE COMMON LAW 31–32 (Mark D. Howe ed., Harv. Univ. Press 1963) (1881).


uniformity in the law of payment systems.9 I am of course not the first to comeupon that problem. The UNPC, for example, rested on the basic premise that thelaw governing issues common to multiple payments should be as uniform aspossible. As the Reporter explained in his memorandum justifying the project,the goal was “to arrive at a set of comprehensive rules applicable in somerespects to all payment systems.”10 Conversely, Peter Alces’s perceptive analy-sis of that project argues that the UNPC went too far by ignoring importantdifferences between payment systems that justify differing rules for devices,primarily credit cards, that provide for the extension of credit.11 More recently,Clay Gillette has presented a particularizing argument about rules for unautho-rized transactions, contending that those rules should turn on the relative abilityof courts and legislatures to identify optimal risk-bearers.12

Looking back from the vantage point of the twenty-first century, it seemsclear that the basic problem with the earlier proposals is not that they areexcessively uniform or excessively particularizing, but that they have notundertaken to consider why it might be useful to have uniform or particularizedrules on particular subjects. I attribute that blind spot in the existing literature tothe historical accident of the structure of the modern commercial-law curricu-lum.13

Even with the law of payments itself, work that attempts to address broaderpolicy concerns is hampered by the balkanized nature of the existing regulatoryapparatus. The UCC itself is promulgated by the ALI and NCCUSL for adop-tion by the various state legislatures. The Federal Reserve—which is motivated

9. The division between the law of sales and the law of payments is a foundational element of ourframework for teaching and writing about commercial law. Uniform Commercial Code (UCC) Article 2governs sales. It is a subset of contract law, traditionally taught either in the first-year contracts course,as a free-standing upper-division course, or as the first segment in a commercial transactions course.Payments are governed by a number of separate UCC articles (3, 4, 4A, and 5), as well as a variety offederal statutes (the EFAA, see Expedited Funds Availability Act, 12 U.S.C. §§ 4001–4010 (2000); theCheck 21 Act, see Check Clearing for the 21st Century Act, Pub. L. No. 108-100, 117 Stat. 1177(2003), to be codified at 12 U.S.C. §§ 5001–18; TILA, see Truth in Lending Act, 15 U.S.C. §§1601–67(f) (2000); and the EFTA, see Electronic Fund Transfers Act, 15 U.S.C. §§ 1693–93r (2000),being the most important). As a course, it is a separate and freestanding division of commercialtransactions, which can be taught on its own or taught as a separate part of a commercial transactionscourse. Aside from notable polymaths like Jim White, leading academics do not often make substantialcontributions to both subjects. I am not sure of another example, but Jim White certainly is a good one.He served as the reporter for the recent revisions to UCC Article 5 (on letters of credit) and (aside fromhis work on his hornbook with Bob Summers) also has written a number of noteworthy articles on UCCArticle 2. See, e.g., James J. White, Autistic Contracts, 45 WAYNE L. REV. 1693 (2000); James J. White,Default Rules in Sales and the Myth of Contracting Out, 48 LOY. L. REV. 53 (2002); James J. White,Good Faith and the Cooperative Antagonist, 54 SMU L. REV. 679 (2001).

10. SCOTT, supra note 4, at 252.11. Alces, supra note 5, at 112–15.12. Clayton P. Gillette, Rules, Standards, and Precautions in Payment Systems, 82 VA. L. REV. 181

(1996).13. I do not attribute the inconsistencies in payment law itself to the law school curriculum. My

point here is only that the gap in the scholarly literature reflects the way in which the curriculum isorganized.


primarily by concerns about stability and to a lesser degree by concerns aboutcost-effectiveness—implements most of the relevant provisions of the EFAA,TILA, and the EFTA. Even the Federal Trade Commission has a minor role,with some frankly protective regulations related to holder-in-due-course sta-tus.14

The basic problem is that payments policy needs to attend more consciouslyto the contexts of the transactions in which payments are made. Existing lawarticulates rules that are bounded almost entirely by the nature of the technologywith which the payment is made. Thus, we have separate rules for wiretransfers, letters of credit, checks, electronic transfers, and the like. That type ofboundary makes sense only for issues driven by the nature of the technology. Itmakes no sense for issues that should be resolved by reference to the nature ofthe underlying transaction in which the payment is made.

At its heart, payments law must resolve four fundamental questions:15 whobears the risk of unauthorized payments; what must be done about claims oferror; when payments are completed so that they discharge the underlyingliability; and when they can be reversed. The first three questions are categori-cally different from the last because they often should be resolved based on thenature of the underlying technology. Thus, for example, with respect to the riskof unauthorized payments, the fundamental question is how to design a systemthat gives adequate incentive to the user to avoid and mitigate losses fromunauthorized transactions, while giving adequate incentive to the system opera-tor to make advances in technology and system design that can avoid andmitigate those losses. In our legal system, we have taken the view for mosthigh-technology payments that an almost complete allocation of the risk ofthose losses to the system operator is appropriate.16

The premise of those rules is that even a complete allocation of loss to thenetwork operator will leave the consumer a sufficient incentive to attend tocontract provisions that resolve the legal questions summarized above. Thatcould be true because of the hassle of reversing unauthorized charges, becauseof doubts that financial institutions readily will fulfill their obligations in such asituation, or even because of ignorance of the legal protections for unauthorizedtransactions. At the same time, the rules reflect the implicit premise that lossesin technology-driven systems are most effectively reduced by technological and

14. See infra note 35 and accompanying text. The minor role of the FTC and the greater role of theFederal Reserve at least in part reflect a congressional compromise designed to ensure that TILA wouldnot be enforced with excessive vigor. See Edward L. Rubin, Legislative Methodology: Some Lessonsfrom the Truth-in-Lending Act, 80 GEO. L.J. 233 (1991).

15. There is some necessary arbitrariness in providing any list of fundamental issues, because any ofthese issues could be divided into multiple parts and because other issues that I omit might be regardedas fundamental. I think, however, that this list is sufficiently general and inclusive to serve the point thatI wish to make here.

16. See Electronic Fund Transfer Act § 909, 15 U.S.C. § 1693q (2000); see also UNIF. NEW

PAYMENTS CODE § 200 (Permanent Editorial Bd. for the Uniform Commercial Code, Draft No. 3, (1983)(general exoneration for unauthorized orders).


system-design initiatives that are exclusively within the control of the systemoperator. Thus, we are not surprised to see major investments in fraud-prevention technology in the credit-card and debit-card sectors. Because thejustifications for those rules relate to the nature of the technology, it is plausiblefor federal law to prescribe such a rule for all electronic transfers from con-sumer accounts. It is less plausible to include a similar rule for credit cardtransactions based on the availability of credit in the transaction.17 It would bemore sensible for that rule to be justified by the fact that the transactions areprocessed and cleared in an electronic way,18 which justifies rules like thosediscussed above.

Rules related to error are similar. The types of events that are likely to lead toan error, as well as the mechanisms for detecting, confirming, and responding toan error, usually depend on the technology that is used to clear and processpayments. Thus, it makes some sense that the rule for transactions processedelectronically, which are covered by the EFTA,19 would differ from the rule fortransactions processed entirely by paper/conventional check transactions, whichare governed by Article 4.20 At the same time, the continuing shift of checktransactions from paper to electronic processing, perhaps to be accelerated bythe Check 21 Act, might undermine that distinction.

Rules that determine when a payment is made21 are similar, in that they arefor the most part made based on the practicalities of a particular system. Thus,in the wire-transfer system, we say that the payment is complete when thebeneficiary’s bank becomes obligated to pay the beneficiary.22 In the checkingsystem, we say that the payment is not complete with respect to an ordinarycheck until the check is paid,23 but that it occurs with respect to a cashier’scheck when the payee accepts the instrument.24

Rules related to reversibility, however, are completely different in that theyshould depend on the dynamics of the underlying transaction in which thepayment is made. In the simplest cases, payment systems are specialized for usein particular situations. Thus, for example, in business transactions, parties oftenchoose to make payments with letters of credit or wire transfers. Those systemsinclude particular rules designed for the particular transactions in which they are

17. See Truth In Lending Act § 133, 15 U.S.C. § 1643 (2000).18. See MANN, supra note 1, at 113–14.19. Electronic Fund Transfer Act § 908, 15 U.S.C. § 1693f (2000). The UNPC adopted a process

similar to the EFTA process. UNI. NEW PAYMENTS CODE §§ 300–06 (Permanent Editorial Bd. for theUniform Commercial Code, Draft No. 3 (1983).

20. There is no specific provision for errors in the checking system. For a general discussion, seeMANN, supra note 1, at 85–96.

21. As suggested above, I mean by this concept the moment when the payment discharges theobligation for which the payment is made. This is distinct, of course, from the concept of finality inArticles 3 and 4, which occurs when the payor bank finally becomes obligated to pay. See U.C.C. §4-215(a).

22. See U.C.C. § 4A-506.23. See U.C.C. § 3-310(b)(1).24. See U.C.C. § 3-310(a).


used, which determine the timing and circumstances in which payments can berecovered or stopped once the process has been initiated.25 Because thosesystems are quite specialized, the system-specific rules work well for them.

It is important to see that the rules make sense because of the underlyingtransaction and not because of anything about the payment instrument itself. Forexample, there is nothing inherent in the use of a bank’s written commitment topay that calls for the formalistic emphasis on both an absolute obligation ofpayment upon presentation of conforming documents26 and at the same time anutterly unconstrained right to refuse payment upon presentation of nonconform-ing documents.27 On the contrary, that structure has grown up solely as anadjunct to the particular sales transaction for which the instrument is commonlyused. If the law of letters of credit makes sense—and for the most part I think itdoes—it makes sense only in light of a practical assessment of the realities ofthe sales transactions in which that law is brought to bear.28

The law of wire transfers is animated by an even more conclusive rejection ofreversibility.29 From the perspective articulated here, that emphasis reflects adesire to create an entirely “pure” payment system, divorced from any transac-tion; the wire transfer is suitable for cases in which the party making thepayment is willing to forgo any payment-related right of recovery at all. Oncethe payment is made by wire transfer, there is no substantial recourse inside thesystem. That makes sense in context, because wire transfers are used typicallyby reasonably informed businesses that select such a pure system in contexts inwhich the most important aspect of the transaction is to provide reliably finalpayment as promptly as possible.30

When we turn to less specialized payment systems, however, the issuesbecome considerably more difficult. Historically, if not in current practice, themost prominent is the negotiable instrument. The most distinctive feature of thenegotiable instrument is the ability of those that acquire the instrument to obtainholder-in-due-course status.31 As a practical matter, that status involves anability to separate the instrument from the transaction as much as possible andthus make the obligation to pay irreversible at an early point, at least as regardsclaims related to the underlying transaction.32

The complicating features of negotiable instruments law, however, largelyoperate to render that separation irrevocably permeable. For present purposes,

25. The rules for letters of credit appear in U.C.C. art. 5 and in the UNIFORM CUSTOMS AND PRACTICE

FOR DOCUMENTARY CREDITS (ICC Publication 500). The rules for wire transfers appear in U.C.C. art. 4A.For secondary discussion, see MANN, supra note 1, at 227–28, 292–98.

26. See U.C.C. § 5-108(a); UNIF. CUSTOMS & PRACTICE FOR DOCUMENTARY CREDITS art. 9.27. See U.C.C. § 5-108(a); UNIF. CUSTOMS & PRACTICE FOR DOCUMENTARY CREDITS art. 13.28. See Ronald J. Mann, The Role of Letters of Credit in Payment Transactions, 98 MICH. L. REV.

2494 (2000).29. See MANN, supra note 1, at 227–28.30. See U.C.C. Article 4A prefatory note.31. See U.C.C. § 3-302.32. See U.C.C. §§ 3-305, 3-306.


what is most important is that the policy justifications for those complicatingfeatures uniformly relate to concerns about the balance of power in the underly-ing transaction for which the instrument was issued. For example, severalarbitrary formalities limit the use of the negotiable instrument to cases in whichthe parties are sufficiently sophisticated and focused to ensure that the paymentinstrument is drafted in a stripped-down form that includes the requisite formallanguage and omits any substantial discussion of the underlying transaction.33

Similarly, even if the instrument is issued in a proper form and transferred in theappropriate way, certain defenses will remain valid against the purchaser. Thesedefenses, the so-called real defenses, address such matters as contracts withminors or contracts procured by fraud; they plainly are designed to protectfundamental concerns about fairness in the underlying transaction.34 Finally, innonmortgage credit transactions that involve consumers, holder-in-due-coursestatus is generally prohibited as a matter of supervening federal law.35

The negotiable instrument, of course, has been superseded for the most partby its main surviving descendant, the check, an instrument for which the classicrules of negotiability have little continuing significance.36 Because the check isless specialized than the letter of credit or the wire transfer or the negotiableinstrument in its heyday, its rules do not reflect the close accommodation to thebalance of the underlying transaction that typifies the law of those earlier,primarily business-related payment systems. Many of the most important rulesin the checking system allocate losses from unauthorized transactions and risksof errors related to the payment device that have little or nothing to do withproblems in any underlying transaction.37 Of course, the focus of modern checklaw on such questions, to the exclusion of any substantial concern for theconsumers that use them, is the basis of many of the most forceful criticisms ofArticles 3 and 4 as they now appear in the UCC.38 But even the checkingsystem includes rules that address the basic problem at the intersection of everypayment system and the transactions for which it is used: the consequences of

33. See U.C.C. § 3-104; see also MANN, supra note 1, at 400–12.34. See U.C.C. § 3-305(a)(1).35. See 16 C.F.R. § 433 (2004); MANN, supra note 1, at 441–42; Ronald J. Mann, Searching for

Negotiability in Payment and Credit Systems, 44 UCLA L. REV. 951, 967–68 (1997)36. See Mann, Searching for Negotiability in Payment and Credit Systems, supra note 35, at

985–1004.37. See MANN, supra note 1, at 69–83, 85–104.38. See, e.g., Robert K. Rasmussen, The Uneasy Case for the Uniform Commercial Code, 62 LA. L.

REV. 1097 (2002). As this Essay suggests, I am not persuaded by much of the criticism. I recognize thatit is difficult to draw a coherent line between commercial and consumer instruments. See, e.g., WilliamJ. Woodward, Jr., “Sale” of Law and Forum and the Widening Gulf Between “Consumer” and“Nonconsumer” Contracts in the UCC, 75 WASH. U. L.Q. 243 (1997). And I think the law does moreharm by watering down rules for transactions between businesses to protect consumers, largely becauseof my sense that the statutes end up including bad rules for businesses that do not in any event provideadequate protection for consumers. Thus, although this is not the place to defend that view, I think it isbetter as a matter of design to articulate separate rules for a describable class of consumer transactions,as federal regulatory legislation traditionally has done in this context.


the payee’s failure to perform. On that point, the UCC grants the check-writer aright to stop payment without any assessment of the validity of the claim.39

The check, however, is now outdated. As we now know, it has been decliningin use for some time.40 The pressure to revise rules related to the check thus willcontinue to decrease. At the same time, consumer use of credit and debit cardsis increasing rapidly.41 Moreover, of importance for our purposes, credit anddebit cards over the last decade have come into dominance in areas in whichthey were not frequently used in the past. In particular, credit cards have cometo dominate payments in remote-purchase transactions, especially on the Inter-net.42 Debit cards, reaching broad use in this country only in the last decade, arenow commonly used in face-to-face transactions and perhaps soon will be amajor option for remote transactions as well.43

Thus, if there is an area of payments law that is both important and currentlycontestable, it is the law that addresses card-based payment transactions. Theremainder of this Essay examines that law in some detail, comparing andcontrasting it to the law of checks, always with a view to discerning the policiesthat motivate the existing structure and explaining how it could be reconstructedin accordance with the framework discussed above.

II. THE EXISTING FRAMEWORK

A. REVERSIBILITY IN EXISTING PAYMENT SYSTEMS

As suggested above, the most contested policy question for card-basedpayment systems is the question of reversibility: when does the consumer losethe right to retract the payment from the merchant? Reversibility also is theissue for which my analytical approach of tying the payments question tounderlying sales policy has the most pronounced application. The rules relatedto errors and unauthorized transactions may differ in minor details for creditcards and debit cards, but for the most part they are quite similar.44 Further-more, as suggested above, I believe those rules generally yield plausibly correctanswers for reasons that do not warrant extended discussion.

For reversibility, however, the situation is quite different. On that topic, the

39. U.C.C. § 4-403. For an argument that this right is unduly narrow, see Alces, supra note 5, at 95.40. The Use of Checks and Other Noncash Payment Instruments in the United States, FED. RESERVE

BULL., Aug. 2002.41. See MANN, supra note 6.42. PAYMENT INSTRUMENTS AS A PERCENTAGE OF TOTAL ECOMMERCE, at http://www.epaynews.com/

statistics/transactions.html#39 (last visited Mar. 11, 2004) (reporting 81.3% for 2002).43. See infra note 98.44. See Truth in Lending Act § 133, 15 U.S.C. § 1643 (2000) (limiting liability of credit card holder

for unauthorized transactions to fifty dollars); Electronic Fund Transfer Act § 909, 15 U.S.C. § 1693(g)(2000) (limiting liability of debit card holder for unauthorized transactions to fifty dollars, unless thecardholder fails to report either the theft of the card or unauthorized transactions that appear on thecardholder’s statement); see also Ronald J. Mann, Regulating Internet Payment Intermediaries, 82 TEX.L. REV. 681 (2004).


paradigm starts from the simple rule for cash payments: if a consumer pays withcash, the “payment” is final at that moment, in the sense that the consumercannot recover the cash. Of course, the consumer might obtain a separate rightto payment from the merchant by establishing some separate claim under thecontract in question. That is, however, quite a different thing from a right toretract the payment itself. For cash payments, such a right of retraction obvi-ously is impractical.

For non-cash payment systems, however, the situation is considerably morecomplicated: the different non-cash payment systems afford consumers differ-ing rights to retract a payment made to a merchant without first proving a failureof the merchant’s entitlement to payment under the contract in question. Thatright of retraction and the system-to-system differences in its application is thefocus of the remainder of this Essay.

Although the landscape is changing rapidly, three systems now dominatenon-cash consumer payments in the United States. As of 2002 (the last year forwhich complete statistics are available), checks still were delivered in 23% ofall retail payment transactions, credit cards were used in about 17% of thosetransactions, and debit cards were used for 11% of transactions.45 The followingsections discuss the reversibility rules that apply in each of those three systems.

1. Reversibility with Checks

The checking system offers the customer what seems at first to be thebroadest right to retract payment, a right to stop payment set out in U.C.C. §4-403(a).46 The customer’s right to stop payment is absolute: the customer canact for any reason it wishes or even, it seems, for no reason at all. The generalconcept, as the comment to that provision explains, is that the right to stoppayment is a basic right that should be incident to a bank account, even if thebank occasionally incurs some loss through the customer’s exercise of thatright.47

However salutary that policy might sound in the abstract, technological

45. See Consumer Payment Systems, NILSON REP., Nov. 2003, at 1, 6. Cash was used in 42% of thetransactions. All other systems (including money orders, prepaid cards, travelers cheques, food stamps,and others) collectively amount to about 7%.

46. That provision states: “A customer . . . may stop payment of any item drawn on the customer’saccount . . . by an order to the bank describing the item . . . with reasonable certainty received at a timeand in a manner that affords the bank a reasonable opportunity to act on it . . . .” A few states (includingNew York) have not yet adopted the current version of Article 4. See Table of Jurisdictions WhereinCode Has Been Adopted, 2B U.L.A. 1 (Supp. 1999–2000) (table indicating that the only U.S.jurisdictions that have not adopted the revised Article 4 are New York, South Carolina, Delaware, andthe Virgin Islands). The older version still in effect in those states grants a substantially identical right tostop payment in U.C.C. § 4-403(1) (pre-1990 version).

47. U.C.C. § 4-403 cmt.1; see Robert D. Cooter & Edward L. Rubin, A Theory of Loss Allocationfor Consumer Payments, 66 TEX. L. REV. 63, 118–19 (1987). The nature of the losses that the commentexpects the bank to bear is obscure, because the UCC plainly contemplates that banks will be permittedto charge customers that choose to exercise the privilege of stopping payment on their checks. SeeMANN, supra note 1, at 14 (discussing UCC provisions indicating a hands-off attitude to regulation of


developments have so overtaken it that in the contexts most important toconsumers it has become almost a dead letter. The difficulty is in the last clauseof the relevant sentence of Section 4-403(a), which limits the time within whichthe customer can send a stop-payment order. Specifically, the notice must besent at a time that gives the bank a reasonable opportunity to act “before anyaction by the bank . . . described in Section 4-303.” That section, in turn,describes the points in the life of a check after which a third party cannotprevent a bank from paying a check; the key point for this discussion is themoment when “the bank settles for the item without having a right to revoke thesettlement under statute, clearing-house rule, or agreement.”48 Thus, the custom-er’s right under Section 4-403 to force its bank to stop payment on a check thatthe customer has written terminates when the bank’s obligation to pay thecheck, normally to a depositary bank or some intermediary that has transmittedthe check to the customer’s bank, becomes final under Section 4-303.49 Interest-ingly enough, the UNPC adopted that rule as the test for all of the paymentsystems that it covered,50 rather than the more generous rules from TILA andthe EFTA discussed below.51

The problem for the customer is that modern settlement mechanisms arelikely to get the check to the customer’s bank, the “payor” bank in the UCC’sterminology,52 quite swiftly. For example, if the customer writes the checklocally, that is, in the metropolitan area in which the customer’s bank is located,the payor bank often will receive the check on the very day that the customerwrites it. Because the federal Expedited Funds Availability Act (EFAA) gener-ally requires depositary banks53 to allow their customers second-business-dayavailability for funds deposited by local check,54 banks into which such checks

checking-account fees); see also Alces, supra note 5, at 95 (discussing how bank’s right of subrogationunder UCC § 4-407 undermines the practical value of the right to stop payment).

48. U.C.C. § 4-303(a)(3).49. The UCC permits an even shorter deadline at “a cutoff hour no earlier than one hour after the

opening of the next banking day after the banking day on which the bank received the check and nolater than the close of that next banking day or, if no cutoff hour is fixed, the close of the next bankingday after the banking day on which the bank received the check.” U.C.C. § 4-303(a)(5). It appears,however, that banks do not generally rely on that provision. See, e.g., Telephone Interview with JoeDeKunder, Vice President, Bank of America, N.A., at 1 (Aug. 23, 1999) (transcript on file with author)(explaining that Bank of America’s policy generally permits customers to stop payment until the bankhas become obligated to pay the check under applicable rules of the clearinghouse or other arrangementunder which the bank receives the check).

50. The definition of “order” in the UNPC is quite broad, and generally would cover what we nowwould regard as checks, ACH transfers, wire transfers, credit card, and debit card transactions, with theodd exception of store cards. UNIF. NEW PAYMENTS CODE § 10 & cmt. 10 (Permanent Editorial Bd. forthe Uniform Commercial Code, Draft No. 3 (1983).

51. See id. § 420. The UNPC did afford a right to reverse orders for three business days, but the rightto do so could and presumably normally would have been waived in the customer’s account agreement.Id. § 425(2).

52. See U.C.C. § 4-105(3) (defining “[p]ayor bank”).53. The depositary bank is the bank into which the merchant deposits the check. See U.C.C. §

4-105(2) (defining “[d]epositary bank”).54. See 12 C.F.R. § 229.12(b) (2004); MANN, supra note 1, at 24–25.


are deposited have powerful incentives to process those checks rapidly. Respond-ing to that incentive, large banks in major metropolitan areas have developedefficient local clearinghouses that typically get local checks into the hands ofthe payor bank by the middle of the night on the date on which those checks aredeposited.55 To protect depositary banks against the risk that their depositorswill withdraw funds purportedly deposited by check before the depositary bankscan learn which checks will be dishonored,56 the clearinghouses include draco-nian rules requiring payor banks to provide swift notice of dishonor, normallyno later than early in the afternoon of the next business day, usually, the dayafter the check is written.57

This acceleration of check processing is an improvement to the checkingsystem, because it reduces the inefficiency of the time lag between the momentthe check is tendered and the moment that the payee reliably has access to thefunds. But the statutory rule that ties the customer’s right to retract/stoppayment to the processing system has important consequences for the consumer.To understand those consequences, consider a typical check, written at ten inthe morning to purchase a lawnmower at a store in the consumer’s hometown,in which the consumer’s bank is located. The merchant might deposit the checklate that afternoon, in which case it probably would reach a clearinghouse thatevening and the payor bank sometime in the middle of the night. The payorbank, in turn, would become finally obligated to pay the check if it did notdishonor the check by early the following afternoon. Thus, the consumer’s rightto stop payment on the check would last little more than twenty-four hours.58

To be sure, many consumers purchase things in locations far from theirhomes, in locations from which depositary banks would not have access tothose speedy local clearinghouses. However, even for remote purchases, the bigpicture for the consumer is not particularly sanguine. For one thing, merchantsare much less likely to accept a check from a customer whose bank is in aremote location than they are to accept such a check from a local customer witha local bank. Moreover, when the merchants do take such checks,59 their banks

55. See MANN, supra note 1, at 43–48 (discussing those clearinghouses).56. Depositary banks are at risk for two distinct reasons. First, they are obligated to release the first

$100 of funds deposited by check on the first business day after the check is deposited, even though it ismost unlikely that they would have received a notice of dishonor by the beginning of that business day.See 12 C.F.R. § 229.10(c)(1)(vii) (2004). Second, they might and often do have funds availabilitypolicies more generous than those required by the statute. Indeed, recent surveys suggest that asubstantial minority of banks, especially large metropolitan banks, have such policies. See BD. OF

GOVERNORS OF THE FED. RESERVE SYSTEM, REPORT TO THE CONGRESS ON FUNDS AVAILABILITY SCHEDULES

AND CHECK FRAUD AT DEPOSITORY INSTITUTIONS 39 tbl.10.1 (Oct. 1996) (on file with author) (reportingFederal Reserve data indicating that about 36% of banks offer same-day availability for local checks).

57. See MANN, supra note 1, at 52–59.58. See id. (explaining clearinghouse processing and showing how the payment becomes final when

the clearinghouse deadline passes without objection by the payor bank).59. Industry statistics indicate that about one-third of checks deposited at large metropolitan banks

are cleared through local clearinghouses, with two-thirds of those checks to be cleared through other,typically slower mechanisms. See MANN, supra note 1, at 52-53, 59-61. Those statistics do not


have every incentive to move rapidly to clear the checks; funds deposited in theform of nonlocal checks ordinarily must be made available to depositors withinfive business days.60 Thus, both the Federal Reserve and the banking industryhave devoted considerable resources over the last few decades to speed theprocesses by which they clear those checks.61

The result is that a substantial majority of nonlocal checks are received by thepayor banks within two or three business days from the day on which they arewritten.62 Thus, even for nonlocal checks, the customer probably has only threeor four business days within which to stop payment.63 Moreover, for remotepurchases it is important to recognize the heightened likelihood of a deferreddiscovery of dissatisfaction with the purchase. The consumer that purchasessomething while away from home often will not discover a problem with thepurchase until returning home; if the return home is just a few business daysafter the date of the purchase, then, as surely as if the check had been writtenlocally, the consumer’s right to stop payment will lapse before the problem isdiscovered.

In sum, however generous the promise of the UCC’s right to stop payment inchecking transactions seems at first glance, the reality is that American consum-ers have no generally effective practical right to retract payments made bycheck. Once the check is handed to the merchant, the recourse of the consumeras a practical matter is limited to the cumbersome device of a legal actionclaiming some defect in the purchased goods or services.64

2. Reversibility with Credit Cards

The above summary might strike the casual reader as unexceptional. After all,

distinguish between retail consumer transactions and other checking transactions. A wholly anecdotalreview of my recently written checks suggests that those statistics, which aggregate all checks,understate the role of local clearinghouses in retail consumer transactions: it appears that none of thelast 100 checks that I have written were payable to a party located outside my local metropolitan area.

60. See 12 C.F.R. § 229.12(c)(1) (2004); MANN, supra note 1, at 25–26.61. For a brief description of the current industry practices, see MANN, supra note 1, at 51–66.62. The most recent published data indicate that in 1997, 59.9% of nonlocal checks were returned to

the depositary bank within four business days, and 82.8% within five business days. There is everyreason to think that processing is significantly faster now than it was then. But even using those figures,if it takes two business days from the date of receipt at the payor bank to the date the check is returnedback to the depositary bank, then the 82.8% figure represents the number of checks received by thepayor bank within three business days of deposit. Availability of Funds and Collection of Checks, 63Fed. Reg. 69,027, 69,028 (Dec. 15, 1998). Even if the payor bank is getting the check back to thedepositary bank on the first business day after it receives the check (which seems most unlikely), 59.9%of the checks are hitting the payor bank by the third day.

63. The right to stop payment normally will end no later than the first banking day after the bankingday on which the payor bank receives the check because the payor bank’s obligation to pay the checknormally will become final on midnight at the end of that first subsequent banking day. See U.C.C. §§4-104(a)(10), 4-215(a)(3), 4-301(a); MANN, supra note 1, at 50–54 (discussing the steps that a payorbank must take to dishonor a check in a timely manner).

64. For a brief discussion of some of the most obvious reasons why such a remedy is not likely to beadequate, see Ronald J. Mann, Verification Institutions in Financing Transactions, 87 GEO. L.J. 2225,2248–49 (1998).


when a consumer purchases something with cash, a bout of next-day dissatisfac-tion will not give the consumer the right to go back to the merchant and forcethe merchant to return the cash. If the consumer wants to retract the cash fromthe merchant, the consumer will have to convince the merchant of the validityof the complaint or, at least, of sufficient sincerity in the complaint that themerchant will disgorge the cash even if it doubts the validity of the complaint.From that perspective, the limited practical effectiveness of the check-usingconsumer’s right to stop payment is not likely to be upsetting.

For the second major non-cash payment system, however, the situation isquite different. If the hypothetical lawnmower purchaser had used a credit card,it would retain, for a period likely to extend several weeks, two overlappingrights to retract payment from the merchant. Those rights both arise not underthe state-promulgated Uniform Commercial Code, which has little or no applica-tion to credit card transactions, but rather under the Truth in Lending Act(TILA), Title I of the federal Consumer Credit Protection Act.65

The first appears in TILA § 170(a),66 which grants cardholders that makepurchases with a credit card the right to assert against their issuing banks anydefense that they could have asserted against the merchants from whom thepurchases were made.67 In legal substance, Section 170(a) articulates a generalanti-holder-in-due-course rule68 for credit card transactions. Thus, the sectionprotects the cardholder’s defenses even when the claim for payment is trans-ferred from the merchant, from which the purchase was made, to the bank thatissued the credit card.

But the practical effect of the rule is somewhat broader, allowing consumersto retract payment from merchants essentially at will. The key is the internalnetwork rules of the major credit card systems. Under those rules, a card-issuingbank can and typically does unwind, or charge back, any credit card transactionfor which a customer disputes its obligation under TILA. Thus, as soon as thecardholder interposes a claim under Section 170(a), the issuing bank charges the

65. TILA is codified at 15 U.S.C. §§ 1601–1667e (2000). The Consumer Credit Protection Act iscodified at 15 U.S.C. §§ 1601–1693r (2000). For clarity, I adhere in the text of this article to thecustomary practice of referring to specific provisions of TILA by the section number of the ConsumerCredit Protection Act itself rather than by the section number in the codification in Title 15 of theUnited States Code.

66. 15 U.S.C. § 1666i(a) (2000); see Regulation Z, 12 C.F.R. § 226.12(c) (2004) (regulatoryexplication of Section 170(a)).

67. For a general discussion of Section 170(a), see MANN, supra note 1, at 114–18.68. The standard holder-in-due-course doctrine is a rule applicable to negotiable instruments, which

grants certain favored transferees immunity from most defenses to the obligation reflected in theinstrument. See U.C.C. § 3-302 (statutory grant of rights of a holder in due course); MANN, supra note1, at 432–43 (secondary discussion of holder-in-due-course doctrine); JAMES WHITE & ROBERT SUMMERS,UNIFORM COMMERCIAL CODE ch. 14 (5th ed. 2000). The legislative history of TILA indicates that theanalogy to that rule drove Congress in articulating the rule in this area. S. REP. No. 93-278, at 9–11(1973). For a thorough and contemporaneous presentation of the argument that holder-in-due-courseprotections for traditional negotiable instruments supported the TILA rule on reversibility, see RolandE. Brandel & Carl A. Leonard, Bank Charge Cards: New Cash or New Credit, 69 MICH. L. REV. 1033(1971).


transaction back to the bank that processed the transaction for the merchant.That bank, in turn, recovers the funds that it advanced to the merchant for thetransaction in question.69 Thus, if the merchant wants to receive payment for thechallenged transaction, the onus is on the merchant to take some action tosubstantiate its right to payment.70

To be sure, Section 170(a) itself is subject to various limitations. For onething, it does not apply to transactions that occur both outside the state of thecardholder’s residence and more than 100 miles from that residence.71 Foranother, the right expires when the cardholder pays the charge in question.72

Nevertheless, even with those limitations, the right seems much more likely tohave significance to the consumer than the stop-payment right for checksdescribed above. Assuming a four-week credit card billing cycle, the averagebill will not come for at least two weeks after the date of the transaction, and thedate on which payment is due will be at least several days after that. Thus, evenfor cardholders that ordinarily pay their entire bills promptly each month,Section 170(a) affords an effective remedy, at least for local transactions, thatextends for a period of weeks after the transaction.

Moreover, the statute also provides a broader remedy for cases in which themerchant fails to perform entirely, as in the transaction that inspired this Essay.

69. The charge-backs ordinarily are effected by netting out the funds to be recovered by the issuingbank from the merchant against funds for new sales transactions, which flow in the opposite direction(from the issuer to the acquiring bank and from the acquiring bank to the merchant). For a briefsummary of the process, see MANN, supra note 1, at 115–16. For industry explanations, see TelephoneInterview with Michael Butts, Creditcards.com, at 2–3 (Oct. 15, 1999) [hereinafter Butts Interview](transcript on file with the author); Telephone Interview with Steven Klebe, Vice President for PaymentIndustry Alliances, CyberSource Corporation, at 3 (Oct. 19, 1999) [hereinafter Klebe Interview](transcript on file with author).

70. See MANN, supra note 1, at 116; Klebe Interview, supra note 69, at 3.71. I have not been able to locate statistics regarding the share of credit card transactions to which

that limitation applies. The question is complicated by the uncertainty in locating mail-order transac-tions submitted by telephone or Internet. Compare Plutchok v. European Amer. Bank, 540 N.Y.S.2d135, 137 (Dist. Ct. 1989) (transaction occurs at the merchant’s location), with In re Standard Fin.Mgmt. Corp., 94 B.R. 231, 238 (Bankr. D. Mass. 1988) (transaction occurs at the consumer’s location).Furthermore, it is not clear how often banks rely on that limitation or how receptive courts would be ifthe banks did rely on it frequently. See Hyland v. First USA Bank, 1995 WL 595861 (E.D. Pa. Sept. 28,1995) (accepting a cardholder’s contention that a bank’s response to a complaint related to an overseascredit card transaction amounted to a waiver of the bank’s right to avoid responsibility based on thelocation of the transaction).

The industry sources to whom I have spoken claim that the Visa and MasterCard networks do notrely upon the Section 170(a) limitation. See Telephone Interview with Paul Confrey, Vice President,Electronic Commerce Planning and Communications, MasterCard, at 7–8 (Nov. 10, 1999) [hereinafterConfrey Interview] (transcript on file with author) (responding “No way Jose” to the suggestion that aMasterCard issuer would reject a claim on that basis); Klebe Interview, supra note 69, at 4–5. Oneexecutive explained: “[T]he cardholders wouldn’t do business with anyone who came up with that.Consumer Reports would write them up within a week.” Confrey Interview, supra, at 8.

72. See Truth in Lending Act § 170(b), 15 U.S.C. § 1666i(b) (2000); Regulation Z, 12 C.F.R. §226.12(c)(1) & n.25 (2004). TILA applies a first-in, first-out rule to allocate payments for the purposeof determining the time at which any particular charge has been paid for purposes of Section 170(a).See Truth in Lending Act § 170(b) (2000); Regulation Z, 12 CFR § 226.12(c)(1) n.25 (2005).


If the case is one in which the goods or services were not delivered at all, thecardholder’s complaint falls within the broad definition of billing errors in TILA§ 161.73 That section is not covered by the geographic limitation in Section 170;thus, it can be applied to transactions of any location. More importantly,billing-error claims need not be presented to the issuing bank for months; thestatute requires only that the cardholder provide written notice to the issuerwithin 60 days after the date on which the creditor sent the relevant statement tothe cardholder.74 Also, the billing-error claim continues even if the charge inquestion already has been paid.75 As with the right to withhold payment underSection 170(a), the ordinary effect of a claim of billing error by a cardholder isa prompt retraction by the issuing bank of the payment that it previously madeto the merchant.76

In sum, the consumer that reaches for the credit card instead of the check-book has a much more effective device for pressing subsequent disputes withthe merchants from whom the consumer buys retail goods and services.

3. Reversibility with Debit Cards

The final consumer payment device of current significance is the debit card.Although the debit card looks much like an ordinary credit card, the cards aredistinct in ways that are quite significant legally, largely because the debit cardauthorizes the merchant and the issuing bank to obtain payment directly from adesignated bank account of the cardholder. Thus, the payment comes from thecardholder’s bank account at the time of the transaction or, at most, a few dayslater; unlike the credit card, there is no delay for the transmittal of a statementand payment of the monthly bill by the cardholder.77

Like the credit card, the consumer protections for debit cards come fromfederal law, in this case the Electronic Fund Transfer Act (EFTA).78 The rulesunder the EFTA, however, differ starkly from the rules under TILA. Specifi-cally, the EFTA contains no analogue either to TILA Section 170(a)’s right to

73. Truth in Lending Act 15 U.S.C. § 1666 (2000); see Regulation Z, 12 CFR § 226.13(a) (2004)(regulatory explication of TILA’s billing-error rules); MANN, supra note 1, at 121–31 (general discus-sion of TILA’s billing-error rules).

74. See Truth in Lending Act § 161(a), 15 U.S.C. §1666(a) (2000); Regulation Z, 12 CFR §226.13(b) (2004).

75. Neither Section 161 nor the regulatory explication of that provision at Regulation Z, 12 CFR §226.13, states expressly that the right to complain of billing errors continues after payment is made. Butgiven the express limitation of the Section 170(a) right to claims for which payment has not been made,the absence of any such limitation from Section 161 and Regulation Z strongly suggests the conclusionstated in the text.

76. See MANN, supra note 1, at 131.77. For a general description of the mechanics of the system, see MANN, supra note 1, at 141–46.78. The EFTA is Title IX, §§ 902–920, of the Consumer Credit Protection Act, 15 U.S.C. §§

1693–1693r (2000). As with TILA, the text of this article adheres to the common practice of citing thesection numbers of the Consumer Credit Protection Act itself rather than the section numbers of that actas codified in the United States Code.


withhold payment or to the billing-error rules in TILA Section 161.79 Thus,payments made with a debit card are just as final as if they had been made withcash.80

B. THE EXISTING POLICY JUSTIFICATION: LIMITING IMPRUDENT BORROWING

Stepping back from the details of the legal rules discussed above, it is easy todiscern an overarching policy justification for the existing framework, rooted inthe long-standing concerns of the Anglo-American legal system about therationality with which borrowers evaluate credit transactions.81

From that perspective, the existing rules divide consumer payments into twoclasses. The first class includes transactions in which the consumer makescontemporaneous payment: cash, checks, and debit cards. Because the con-sumer at the time of the transaction understands that the payment is being mademore or less immediately, the consumer is treated as having adequately assessedthe wisdom of the payment in question.

Credit cards, however, are quite different from that perspective, because theconsumer that purchases with a credit card does not make immediate payment.Rather, although the merchant receives contemporaneous payment,82 the pay-ment by the consumer is deferred automatically until a statement is receivedand, at the consumer’s option, more or less indefinitely, as permitted by thestrikingly lenient repayment options typical of the modern American creditcard.83

79. The EFTA and Regulation E, 12 CFR § 205, do include procedures for resolving errors, but thedefinition is limited to unauthorized transactions or other errors in posting the transaction to theaccount. Unlike the analogous provisions in TILA and Regulation Z, the debit card rules do not extendto a failure of the merchant to perform its obligations in the transaction. See Electronic Fund TransferAct, 15 U.S.C. § 908(f) (2000); Regulation E, 12 CFR § 205.11(a)(1) (2004).

80. The distinction is not an accident. As Jim Rogers has pointed out to me, consumer advocatespressed this point vigorously in the debates about adoption of the EFTA. Congress simply refused toinclude it. See NAT’L COMM’N ON ELEC. FUND TRANSFERS, EFT IN THE UNITED STATES: POLICY RECOMMEN-DATIONS AND THE PUBLIC INTEREST 49–53 (1977) [hereinafter EFT IN THE US REPORT] (recommendingagainst a rule of reversibility for electronic fund transfers). Still, it is difficult to explain the differingrules as a positive matter. Cf. Gillette, supra note 12, at 207–09 (noting the oddity of similar rules forcredit cards and debit cards in an area where they seem to be functionally dissimilar). The principalarguments advanced by the National Commission on Electronic Fund Transfers to support its rejectionof the rule were that too many merchants would refuse an electronic payment system with reversibilityand that consumers who wanted reversibility could use checks or credit cards. EFT IN THE US REPORT,supra, at 52–53. On the first point, the success of credit cards is a strong counterexample. I have neverseen an American merchant that accepts debit cards but not credit cards. On the second point, thediscussion above explains that checks no longer provide a substantial reversibility option. Finally, as Idiscuss in Part IV, credit cards provide such an option only for those wealthy enough to have them.

81. Peter Alces argues that credit and debit transactions should be treated differently, but he does notclearly make the link to behavioral concerns that I use here to justify that distinction in existing law. SeeAlces, supra note 5.

82. See MANN, supra note 1, at 115–16.83. The perception that those options are too lenient has motivated congressional efforts to require

specific disclosures regarding the length of time repayment would take at the minimum payment rates.See Dean Anason, LaFalce Sees Compromise as Reform’s Best Hope, AM. BANKER, Apr. 29, 1999, at 3


The Anglo-American legal system has a long tradition of protecting borrow-ers from the folly of imprudent borrowing. The most famous example surely84

is the centuries-long effort of English courts to invalidate a series of creditordevices that had the effect of granting mortgage creditors a broad right to takereal-estate collateral from borrowers that failed to perform precisely as they hadpromised at the time of the loan.85 That instinct continues to have broadapplication today, as courts steadily broaden the range of devices to which thatinvalidation rule extends.86

Although it is a bit much to superimpose the insights of modern academicliterature onto the policy instincts of the medieval English judiciary, that policyfinds broad support in the specific concerns of the nascent behavioral economicsmovement. Scholars in that field often point to an overlapping set of tendenciesthat generally lead a normal individual to underestimate the likelihood ofnegative future events that have not previously been salient in the experience ofthe individual’s circle of personal acquaintances.87 A similar, related phenom-enon leads to systematic underestimation of the likelihood that a negative eventwill happen to the estimating individual, even if the individual accuratelyunderstands the overall likelihood of the event.88 Both of those phenomena areexacerbated by the likelihood that consumers have higher discount rates forevents perceived as likely to occur far in the future than they do for eventslikely to occur in the immediate future.89 Thus it is reasonable to worry thatborrowers entering into credit transactions do not adequately weigh harmslikely to befall them from the difficulties that might come at the time that

(discussing possible disclosure requirements); Dean Anason, Bankruptcy Bill Is Getting Last-MinuteTweaks, AM. BANKER, Sept. 10, 1999, at 2.

84. At least from the perspective of an academic who began his career teaching in the area ofreal-estate finance.

85. See RESTATEMENT OF MORTGAGES § 3.1 cmt. a (1997); Ann M. Burkhart, Lenders and Land, 64MO. L. REV. 249, 250–67 (1999). For citations to more detailed treatments, see 1 G. NELSON & D.WHITMAN, REAL ESTATE FINANCE LAW § 3.1, at 32 n.1 (1993) (Practitioner Treatise Series).

86. See, e.g., RESTATEMENT OF MORTGAGES, supra note 85, §§ 3.1–3.4 (articulating rules for modernreal-property applications); WHITE & SUMMERS, supra note 68, § 21.3 (discussing the application ofsimilar rules to transactions that purport to involve leases of personal property).

87. See, e.g., Amos Tversky & Daniel Kahnemann, Judgment Under Uncertainty: Heuristics andBiases, 185 SCIENCE 1124, 1127–28 (1974) (discussing availability biases).

88. See, e.g., Laurie Larwood & William Whittaker, Managerial Myopia: Self-Serving Biases inOrganizational Planning, 62 J. APPLIED PSYCHOL. 194 (1977); Ola Svenson, Are We All Less Risky andMore Skillful Than Our Fellow Drivers?, 1981 ACTA PSYCHOLOGICA 143; Neil D. Weinstein, UnrealisticOptimism About Future Life Events, 39 J. PERSONALITY & SOC. PSYCHOL. 806 (1980) (discussing theoptimism bias).

89. The phenomenon generally is known as hyperbolic discounting. For a discussion in this context,see Stefano Dellavigna & Ulrike Malmendier, Contract Design and Self-Control: Theory and Evidence,119 Q.J. ECON. 353 (2004). See generally Shane Frederick et al., Time Discounting and TimePreference: A Critical Review, 40 J. ECON. LIT. 351, 360 (2002); Jonathan Gruber & Botond Koszegi, IsAddiction “Rational”? Theory and Evidence, 116 Q.J. ECON. 1261 (2001); George Lowenstein &Richard H. Thaler, Anomalies: Intertemporal Choice, J. ECON. PERSP., Fall 1989, at 181, 184–87(discussing dynamic inconsistency in discount rates); Richard H. Thaler, Some Empirical Evidence onDynamic Inconsistency, 8 ECON. LETTERS 201, 202 (1981).


repayment is due.90 Also, with a little more subtlety, those that pay in advancemight underestimate the likelihood that they will harm their strategic relationswith the merchant by agreeing to pay now and receive the subject merchandiseor services later.

Finally, that concern works well as a lens for understanding TILA. TILA doeslittle or nothing to regulate anything about credit card transactions divorcedfrom their credit-related features.91 The bulk of its regulations appear to rest onthe concern that crafty credit card issuers will trick consumers into using creditcards without understanding the costs of repayment.92 Thus, the statute reliesheavily on a variety of disclosure rules reflecting the well-intentioned notionthat mandatory disclosures will resolve the cognitive problems that afflict thepotential credit card user.93

III. A BETTER FRAMEWORK: REDRESSING AN IMBALANCE OF LEVERAGE

Although the concern about consumers’ general failure to appreciate the riskof borrowing transactions seems to provide a relatively coherent normativeexplanation of the current system, a comparison of that explanation to theframework discussed in Part I suggests that the existing framework is deeplyflawed. Three general points illustrate my concerns. First, developments inpayment systems render the distinction between credit and debit paymentsillusory. Second, although the point necessarily is subjective, and not to bepressed absolutely, I believe that the concern about cognitive problems in credit

90. In addition to the academic studies that generally relate to discounting problems (several ofwhich are cited supra note 89), there is some limited empirical evidence to support the existence of theproblem in consumer credit markets. See Joshua Brockman, Survey by Fannie Says People Underesti-mate Effects of Poor Credit, AM. BANKER, July 30, 1999, at 14 (reporting study by Federal NationalMortgage Association indicating that consumers systematically underestimate the adverse effects ofpoor payment behavior on future lending transactions). Oren Bar-Gill argues vigorously that much ofthe credit card system is driven by issuer exploitation of behavioral problems of consumers. OrenBar-Gill, Seduction by Plastic, 98 NW. U. L. REV. 1373 (2004). Because his argument focuses on overallinterest-rate pricing and efforts to cause consumers to accept cards rather than the actual purchasingtransactions it is not directly relevant to the issues that I address here. As I explain in Ronald J. Mann,Credit Cards, Consumer Credit, and Bankruptcy (2005) (unpublished manuscript), I am impressed withhis argument that behavioral concerns have a significant impact on the design of credit card products. Iam less sure, however, of his empirical conclusion that this aspect of the design contributes significantlyto the policy problems related to credit card use.

91. For a detailed description of the rules imposed by TILA, see 2 BARKLEY CLARK & BARBARA

CLARK, THE LAW OF BANK DEPOSITS, COLLECTIONS AND CREDIT CARDS ¶15.04 (1999).92. A particularly telling example is the odd provision that bars the mailing of unsolicited activated

credit cards. See Truth in Lending Act § 132, 15 U.S.C. § 1642 (2000); Regulation Z, 12 C.F.R. §226.12(a) (2004). The legislative history shows that concern about this practice was common at thetime. See Unsolicited Credit Cards: Hearings on S. 721 Before the Subcommittee on FinancialInstitutions of the Senate Committee on Banking and Currency, 91st Cong., 2d Sess. (1970).

93. For a detailed description of various ways in which TILA fails at that purpose, see Mann, supranote 6.


card transactions is in practice largely irrelevant.94 Third, and most importantly,I think that concerns about an imbalance of leverage in dispute resolution aremuch more significant, and call for an even broader consumer-protection policythan the credit-cognition policy justifies or the existing legal rules provide.

A. THE ILLUSORY CREDIT/DEBIT DISTINCTION

The first point looks to the reality of current market use of payment systems,in which the distinction between debit and credit payment devices has erodedalmost completely from the consumer’s standpoint. The key point here has beenthe recent rise of the debit card from a highly specialized and limited retailpayment instrument to an instrument of widespread use and soon-to-be univer-sal acceptance.

Although the debit card was invented more than thirty years ago,95 it did notrise to prominence until the last days of the millennium. Many factors doubtlessplay a role in that rise, including the steadily decreasing costs and increasingreliability of the dialup telephone connections on which retail debit card usagedepends.96 But one factor is crucial: the development of the PIN-less Visa andMasterCard debit card products (VisaCheck and MasterMoney, respectively).97

Although neither of those products existed until the mid-1990s, they alreadyhave surpassed in market share the traditional PIN-based regional debit cardnetworks that have been operating for much longer. Starting from less than 400million transactions worth only $5 billion in all of 1994, by 2002 the PIN-lesssystems were used in more than 8 billion transactions for $318 billion, com-pared to only 5.2 billion transactions worth $162 billion for the conventionalPIN-based products.98

The cynic might attribute the rapid growth of the PIN-less systems to anunrelenting advertising campaign (especially for the VisaCheck card).99 How-ever, a more fundamental reason for the success is the advantage that the Visaand MasterCard products have over the traditional regional cards at the point ofsale: an installed user base that includes an overwhelming majority of retailmerchants in the United States. Now that credit card transactions are cleared

94. That is not to say that there are no relevant cognitive concerns. On the contrary, I think cognitiveconcerns are a major problem in the credit card industry. See Mann, supra note 6, at 90. It is only to saythat they are not related to the issue of reversibility in any plausible way.

95. See D. BAKER ET AL., THE LAW OF ELECTRONIC FUND TRANSFER SYSTEMS: LEGAL AND STRATEGIC

PLANNING ¶ 7.02 (rev. ed. 1999) (discussing the early history of the use of the debit card at retaillocations).

96. For discussion of that point, see Ronald J. Mann, Information Technology and Non-LegalSanctions in Financing Transactions, 54 VAND. L. REV. 1627, 1630–36 (2001).

97. The following discussion draws heavily on MANN, supra note 1, at 144–46.98. See U.S. General Purpose Cards, NILSON REP., Aug. 1999, at 1, 9–10; U.S. Debit Cards, NILSON

REP., Apr. 2003, at 1, 6–7.99. Early advertisements featuring Bob Dole were particularly notable. See Gary Tuchman, Bob

Dole: Taking Madison Avenue By Storm, CNN, Aug. 7, 1997, at http://www.cnn.com/allpolitics/1997/08/07/dole (describing the success of those advertisements).


almost exclusively by telephone connections,100 the basic infrastructure for thedebit card is in place at almost all Visa and MasterCard merchants. Thosemerchants can take debit card transactions using the same terminals with whichthey take Visa and MasterCard credit card transactions.

The difference in the nature of the authorization transactions—an examina-tion of a particular bank account rather than a particular line of credit—seemsinsignificant to the merchant in the current, telephone-based clearing environ-ment. In either case, the merchant’s terminal dials up the designated telephonenumber and transmits the card number and related information about the cardand the transaction. The issuer’s computer simply checks a different databasefor transactions that use debit card numbers than it does for transactions that usecredit card numbers. The same analysis applies to the collection arrangements,which require deduction from a checking account instead of posting to a creditcard account. That difference is irrelevant to the merchant, which receives fundsin its account promptly after it forwards the appropriate transaction information.Thus the merchant that clears MasterCard and Visa credit card transactionselectronically needs nothing new to accept the VisaCheck and MasterMoneydebit card products.

A regional debit card network attempting to persuade a merchant to accept itscards for point-of-sale transactions has a much more difficult time because themerchant necessarily will have to invest in some additional technology. In somecases, the debit card network may require the use of a different terminal—onethat can accept entry of a PIN—which necessarily will increase costs for themerchant. At a minimum, the merchant will have to alter its existing proceduresto cause its terminal to recognize the situations in which it must dial thetelephone number for the regional debit card network instead of its regularacquiring credit card bank.101 That might sound like a simple task, but it willseem complicated to some merchants, and no alterations will be necessary formerchants that choose to accept only VisaCheck and MasterMoney.

Furthermore, the merchant is quite unlikely to suffer a significant loss ofbusiness because of a failure to accept PIN-based debit cards. Currently, few ofus will enter a store planning to buy, but depart when we discover that the storetakes a Visa or MasterCard product but not our ATM card. The converse, ofcourse, is even more telling. Quite a number of consumers would depart withoutbuying or choose not to return if a store did not accept Visa and MasterCardproducts, even if the store did accept an ATM card from a large local bank.102

100. See Jeremy Quittner, Processing Paper on Wane, but Still Lucrative, AM. BANKER, Nov. 1,1996, at 10.

101. The acquiring bank in a credit card transaction is the bank that processes the transactions on themerchant’s behalf, forwarding them through the Visa and MasterCard network to the various issuers,and providing payment to the merchant after deduction of the appropriate fees. For brief discussion, seeMANN, supra note 1, at 111–14.

102. The ATM card faces an additional disadvantage in those metropolitan areas where major banksare not members of the same network, because a retailer that makes an arrangement with one network


Finally, and perhaps a bit less justifiably, Visa and MasterCard adopted theforceful tactic of insisting that their merchants accept all of their products, debitand credit. Thus, any merchant that already accepts the Visa and MasterCardcredit card products was obligated to accept the Visa and MasterCard debit cardproducts. A number of large merchants were unhappy about that, largelybecause the Visa and MasterCard debit card products exact discount feessubstantially higher than those charged by the traditional regional debit cardnetworks. Retailers responded by instituting antitrust litigation that eradicatedthe practice.103 By the time the tying practice was eradicated, however, the debitcard system already had become widespread.

The result is that the leading credit card products and the leading debit cardproducts in our country now use precisely the same technology. The cards havethe same appearance, they use the same anti-fraud technology, they are clearedthrough the same network, their transactions are verified in the same way—bysignature—and they are issued by the same entities. This fosters a world inwhich the consumer often cannot even tell whether the card pulled from thewallet is a credit card or a debit card.104 Indeed, because many banks issue cardsthat have both debit and credit features, the answer often is that the card is both.In that case, the choice depends on nothing more than which button the clerkpresses on the telephone terminal. Moreover, in some cases, under currenttechnology, a consumer can use a card only by pushing a button at the checkoutcounter that inaccurately identifies the card!105

would gain access only to the customers whose banks use that network. The deep market penetration ofVisa and MasterCard products in terms of both merchant acceptance and consumer usage obviates thatproblem for their products.

103. For discussion of the retailers’ success, see Robert A. Bennett, The Retailers’ Home Run,CREDIT CARD MANAGEMENT, July 2003, at 24.

104. That is not an accident. See Jennifer Kingson Bloom, Visa Stands by Updated Debit Card,Though Banks’ Response Is Cool, AM. BANKER, July 28, 1999, at 1, 13 (emphasizing the effort by Visato obscure the difference between the credit and debit products at the consumer level: “Few consumersknow those details – and this is by design. Consumers ‘won’t know the difference, so there is no pointconfusing them.’”) (quoting Visa’s vice president for debit products); see also Your Debit Card Is theKey To Protecting Your Money, USA TODAY, Oct. 5, 1999, at 14A (full-page advertisement for Visadebit cards, which neglects to mention the disadvantageous EFTA rules discussed in the text). Almostevery year in my Payment Systems course I have the experience of students insisting that they have astrange kind of credit card in which the bank automatically deducts the amounts from their bankaccount every month. The card, of course, is not a strange kind of credit card—it is a debit card—andthe deductions are made daily, not monthly; the confusion on that score arises because the cardholdersreceive statements showing the deductions only on a monthly basis.

More recently, a few American banks (such as Citibank)—adopting the typical practices of banks inJapan and some European countries—have started to offer a product in which the credit card bill is paidautomatically each month by a deduction from the customer’s bank account. Finally, falling in betweenthose products is a delayed debit product in a pilot program from MasterCard, in which all debitcharges are aggregated and posted to an account once or twice a month. See Charles Davis, DelayedDebit Card Differentiates Offering, CARDS INTERNATIONAL, January 2005, at 11.

105. I speak from experience with the Verifone terminal at a prominent Austin merchant. I have aVisaCheck card from my local bank, which can be used either as a PIN-based debit card through a localEFT network or as a PIN-less Visa debit card. Because the merchant in question does not accept


In sum, the practical distinction in modern commerce between a credit cardand a debit card is one that would be obscure to even sophisticated consumers.Given the limited likelihood that the typical consumer understands the distinc-tion, it is at best dubious for the law to place as great a weight on thatdistinction as it currently does.

B. THE LIMITED USE OF CREDIT WITH CREDIT CARDS

The confusing similarity of the leading debit and credit card products wouldnot matter so much if the credit aspect of the transaction were central to theconsumer’s mind. A consumer buying with a credit card would regard thetransaction as substantially less serious and immediate than a transaction with adebit card.

In the modern era, however, the credit aspect of the purchase transaction isincreasingly obscure. For many of us, the credit card is used much more forconvenience than for any purpose related to borrowing. To put it another way,we often pull a credit card from our wallet not because we lack the presentwealth to pay the purchase price for the item in question, but for some otherreason unrelated to a credit decision.

If there is anything that proves that point, it is the rise of the “convenience”credit card user, who charges purchases on the card but pays off the bill in fulleach month.106 Once a relatively small sector of the market, convenience usersnow constitute about 40% of the market.107 For those users that consciouslylimit their spending to what they can repay out of resources available thatmonth, it is almost deceitful to treat the transaction as one involving a signifi-cant extension of credit.

Many people use the credit card for reasons completely unrelated to credit.The most economically rational is a desire to get the “float”—the use of thefunds during the time that elapses between the date of the transaction and thedate that the credit card bill must be paid to avoid a finance charge. But the

PIN-based debit cards, the card will not be accepted if I select “ATM” on the terminal. It will work,however, if I select “credit,” although the transactions are processed as debit transactions, beingwithdrawn from my bank account in a few days. It seems at best ironic that I should be denied TILAprotections for a transaction in which I am forced to present my card as a credit card.

106. Because many American credit card issuers no longer charge annual fees, the profits fromissuing credit cards largely come from interest charges and late fees. Thus, the convenience users, evenif relatively creditworthy, are relatively unprofitable customers for the credit card issuer. See LisaFickensher, New Discover Cards Aimed at People Who Run Balances, AM. BANKER, Apr. 17, 1995, at 1(attributing the limited profitability of the Discover card to its unusually large share of convenienceusers); Lisa Fickensher, Citi’s Card Unit Says It Will Stick to Basics, AM. BANKER, July 27, 1999, at 14(reporting estimates that 80% of industry profits come from those who are not convenience users).

107. It rose steadily from 28.6% in 1990 to 44.4% in 2000, but has fallen each of the last two yearsto 39.9% in 2002. See CARDDATA, BANK CREDIT CARD CONVENIENCE USAGE–CURRENT, available athttp://www.cardweb.com/carddata/charts/convenience_usage.amp (last visited Jan. 13, 2004) (subscrip-tion required). In my work on credit cards, I argue that a more useful metric of the extent of borrowingtracks the ratio of outstanding balances at any given time against the annual credit purchase volume;that figure has fluctuated for the last decade between 55 and 65%. See Mann, supra note 6.


rapid rise of the debit card, which has no such float, suggests that the float is notthe dominating market factor in the choice of a credit card over cash or a check.

More likely reasons probably include the desire not to have to carry enoughcash to cover all of the purchase transactions of our daily life. That is not just adesire to limit the risk of a loss from theft, but also has to do with the relativeease with which credit cards can be used as compared to the ease with which wecan get cash from our banks. Many of us stand at retail counters thinkingsomething like, “If I use a credit card here it means I won’t have to stop at theATM to get money before I go to work tomorrow.”

Another consideration for some relates to the ease of recordkeeping. If mosttransactions can be pushed onto a single credit card, which provides a singleconsolidated monthly statement, the task of monitoring expenses to reviewcompliance with a budget (or, perhaps more commonly, to understand thefailure to comply with a budget) is rendered much easier.108

Still another reason, at least by comparison to the check, is the desire to closethe transaction rapidly. Although this has not always been true, it now plainly isthe case that the check is the slowest of retail payment mechanisms. When agrocery store customer pulls out a checkbook to pay for groceries, the hurriedcustomers in the line behind inevitably sigh, knowing that their wait will beprotracted by the additional time for the consumer to write the check and for theclerk to decide whether to accept it.

Finally, a recent factor is the affinity program. A credit card often carries withit “perks” of some kind that have value to the cardholder. I am motivated toobtain frequent flyer credits on my favorite airline; a colleague is motivated toobtain credits toward purchase of an automobile from a particular manufacturer.Those perks give all of us who have such cards an incentive to place charges ona credit card—especially large charges—even if we could just as easily write acheck, because passing the transaction through the credit card issuer earns ussomething of value that amounts to a discount on the transaction price.109

Those points speak to my thesis with some ambiguity. One could say that thereasons for using a credit card are likely to induce many to purchase things oncredit without realizing that they are borrowing, thus pushing credit card usersinto an imprudent cycle of borrowing. And I recognize, of course, that creditcards are still used to borrow money. Americans as a class now carry astaggering amount of credit card debt, in the range of 750 billion dollars.110

108. American Express was a leader at pushing this feature of the credit card, with its user-friendlysubject-oriented statements.

109. That motivation, of course, can be unprofitable for the credit card company, which earns noincome on such convenience-use transactions. It should surprise nobody that credit card issuers havenoticed that practice and are beginning to respond to it. See, e.g., Antoinette Coulton, Banc One ClipsWings of Convenience Users, AM. BANKER, Aug. 1, 1997, at 12 (reporting program under which BancOne gives lower affinity credits to convenience users than other users). For a thorough discussion ofaffinity programs and their problems, see Mann, supra note 6.

110. See Credit and Debit Cards U.S. Projected Thru 2010, NILSON REP., Sept. 2002, at 7 ($723.66billion outstanding on U.S. credit cards at the end of 2001).


Indeed, I argue in related work that the rise of credit card debt has contributedin part to a disturbing rise in consumer bankruptcy in this country.111

The extent of credit card borrowing does not undermine my point that thedecision to use a credit card is not necessarily the same as the decision toborrow.112 In practical import, two different decisions must be distinguished:the decision to use a credit card to make a purchase and the decision to borrowfrom a credit card issuer. Obviously there are cases in which people purchasethings on credit cards planning to pay for them at the end of the month, but thenfind themselves unable to pay. For those cases, the credit cognition concern hassome weight. However, the existence of some range for the credit cognitionconcern does not respond to my basic view that the credit card transactions raisemajor policy considerations as a payment method, entirely apart from theirborrowing aspects.113

The cognition concern need not be discarded, but just the same, it should notbe pushed to the exclusion of other concerns that have more practical signifi-cance in the broad range of daily transactions. The fact is, for many of us thecredit card involves no credit at all. A sound payments policy must take accountof the dynamics of those transactions as well as the more traditional credittransactions. The next section suggests that the reversibility rule in TILAresponds well to a policy concern common to all of those transactions, and todebit card transactions as well.

C. LEVERAGE AND DISPUTE RESOLUTION

The preceding pages suggest a considerable tension in the existing legalframework. Part II discusses a variety of legal rules that provide protections forconsumers limited for the most part by the time when they pay their credit cardbills. The preceding sections of this Part, in contrast, discuss the likelihood thatmany consumers arrange their affairs so that they take care to pay those billspromptly after receipt, to make sure that they do not incur any charges to thecredit card issuer.114 But the cardholder that diligently pays bills at the same

111. Mann, supra note 90.112. It may be that the situation will change as the debit card grows in popularity and as credit card

issuers increase efforts to repel confirmed convenience users, see, e.g., supra note 109. Those twochanges together suggest the possibility of a market in which different groups of consumers use creditcards and debit cards: those who borrow use credit cards and those who do not use debit cards. Forexample, in a world in which credit card companies charge annual fees and deny affinity benefits toconvenience users, why would those users choose a credit card instead of a debit card that their bankmight provide to them for free (on the premise that it is cheaper for the bank to process debit cardtransactions than it is to process checks). The existing situation is in such flux, however, that it isdifficult to predict the ultimate outcome. For this Essay, however, it is enough to note that such ascenario would not undermine my point: that the debit/credit distinction should be irrelevant to thereversibility question whether or not it is important in practice.

113. For a comprehensive discussion of the policy issues related to borrowing on credit cards, seeMann, supra note 6.

114. The fee-free time is shortening rapidly, as credit card issuers struggle to limit the costs ofservicing convenience users, and shift to a model that relies more heavily on those fees. See Noreen


time blindly gives away many of the valuable rights that TILA grants to thecardholder.

Of course, if you accept the policy premise of TILA then those protectionsmake no sense once the bill has been paid. On the other hand, if you sense awindfall for the merchant that has sold to a convenience user as compared to amerchant that has sold to a credit user, then some broader normative justifica-tion underlies your instincts—a justification that survives the cardholder’spayment to the issuing bank. That justification, I submit, is supplied from theframework articulated in Part I of this essay, by reference to the underlyingsales transaction. In this context, the concern is the imbalance of leveragebetween the typical consumer and the typical merchant.115

1. What Policies Does TILA Serve?

As summarized in Part II, the practical impact of the relevant provisions ofTILA is to alter the burden of going forward in a dispute about a retail purchasetransaction. Without TILA—that is, in a cash or debit card transaction—themerchant has the funds and the consumer can obtain satisfaction only throughthe exercise of some legal remedy or threatened non-legal sanction such aspublic outcry designed to harm the merchant’s reputation. Conversely, if TILAintervenes, then the charge will be removed from the cardholder’s account andthe funds retracted from the merchant, at least temporarily.116 Thus, the mer-chant will have to take action to obtain the disputed funds. With zero transactioncosts and perfectly balanced positions that change would not matter. But thegeneral imbalance in litigating capabilities between merchants and consumers atleast suggests the possibility that the shift of the burden of going forward couldimprove the effectiveness of the system considerably.117

At bottom, the consumer with TILA as a weapon is much better positioned inthe dispute than the consumer without TILA. Moreover, if the improvement of

Seebacher, Consumer Issues: Keep Credit Card Penalties from Imprisoning You, DETROIT NEWS, Sept.13, 1999, available at 1999 WL 3938007 (reporting programs by card issuers to shorten grace periodsand, in some cases, cancel the accounts of convenience users). Credit card issuers have had less successin efforts to enforce late fees more rigorously. See Miriam Kreinin Souccar, Late Fees Seen Backfiringon Card Issuers, AM. BANKER, Oct. 19, 1999, at 1, 18 (reporting three 1999 lawsuits against large cardissuers alleging improper assessment of late charges and other fees). For discussion of some policiesresponsive to that problem, see Mann, supra note 6.

115. Cooter & Rubin allude to that imbalance briefly, but suggest that it has little interest for thembecause it is “better examined as [an] aspect of sales law, not payment law.” Cooter & Rubin, supranote 47, at 121. As this Essay suggests, I think it has much more to do with payment law than they do.

116. The statute requires the issuing bank to cease any efforts to collect the charge from theconsumer until it has conducted a reasonable investigation of the consumer’s allegations and deter-mined that they are unfounded. Truth In Lending Act, 5 U.S.C. § 161(a)(B) (2000); see Regulation Z,12 C.F.R. §§ 226.13(d)(1), (f) n. 31 & (g) (2004).

117. See Cooter & Rubin, supra note 47, at 80–81; 116–17 (discussing reasons why consumers arelikely to underenforce their legal rights); see also Gillian Hadfield, The Price of Law: How the Marketfor Lawyers Distorts the Justice System, 98 MICH. L. REV. 953 (2000) (arguing that the market forlawyers systematically limits the relative availability of lawyers to individuals and expands the relativeavailability of lawyers to commercial enterprises).


the consumer’s position seems appropriate then the lesson of the earlier sectionsof this part is obvious: the improvement in the consumer’s position should notdepend on whether the consumer uses a debit card or a credit card. To put itbluntly, the distinction in consumer protections between credit cards and debitcards rests on a circumstance—the possibility that the consumer will fail to paythe bill at the end of the month—that is irrelevant to the basic transactionbetween the consumer and the merchant. Thus, it should be irrelevant to theprocedures that facilitate the consumer’s ability to obtain satisfaction in thattransaction.

2. Is TILA Beneficial?

It is, of course, difficult to be sure that the protections offered by TILAprovide a net benefit to all consumers.118 If those protections impose costs onmerchants and issuing banks, as they almost certainly do, then it is natural toexpect the charges that merchants and credit card issuers charge to customersand cardholders will rise to compensate for those additional costs. If thoseadditional charges exceed the benefits cardholders obtain from TILA, then theTILA protections impose a net loss on the cardholders for whose benefit theyseem to exist.

The greatest problem in assessing those provisions is the difficulty of obtain-ing the relevant empirical information. In particular, it would be most useful toknow the frequency of cardholder claims under TILA and the frequency withwhich those claims ultimately prevail. If we knew, for example, that such claimsare raised relatively rarely but succeed quite commonly when they are raised,we might conclude that the provisions impose nugatory costs on issuers andmerchants and substantial benefits to the cardholders that use them.

Conversely, if we knew that cardholders frequently interpose frivolous claimssolely to avoid payment of just obligations, we might conclude that the coststhat the provisions impose on the system outweigh any benefits they mightbring to worthy cardholders. The difficulty of evaluating that question isaggravated by the impossibility of drawing firm conclusions from a finding thata large fraction of cardholder claims are rejected. The problem is that it isentirely possible that cardholder TILA claims are rejected not because they aremeritless, but because the same litigation imbalance that could lead merchantsto reject informal inquiries by customers—thus leading the customers to invoketheir TILA rights—would lead the same merchants to reject the claims whenpresented through the issuing bank under TILA.

To the extent any light can be shed on those empirical questions, myimpression based on a few interviews with industry sources is that the formerpicture is closer to the truth in two important ways. First, consumers interpose

118. See Cooter & Rubin, supra note 47, at 117 (suggesting that it “can be only a matter ofspeculation” whether an enhancement in the consumer’s ability to shift losses is worth higher chargesfor consumers in general).


chargeback claims quite rarely. All of the sources to whom I spoke agreed thatthe general rate of complaint-based chargebacks is quite small, significantlybelow one percent.119 That should not be surprising given the obvious likeli-hood that the noncash transaction costs—the hassle of successfully shepherdingsuch claims through the complex TILA process—will exceed the expectedrecovery on any but the largest and most plainly meritorious consumer com-plaints.

Second, the general perception in the industry is that merchants accedequickly to quality-related complaints by their customers.120 As one executiveexplained, “the merchant if they are smart would just immediately issue a creditand not argue at all.”121 To be sure, the willingness to accommodate is likely tovary from merchant to merchant and also differ with the size of the claim.Furthermore, it is not clear that merchants are motivated by a perception of theirown fault. The apparent motivations are much broader and induce concessionby the merchant without regard to the merits of the consumer’s claim. Amongother things, the system charges merchants a fee for each transaction thatconsumers charge back.122 Also, more tellingly, banks charge higher discountrates—higher charges on all transactions, not just disputed ones—to merchantsfor whom an unusually high rate of transactions are charged back.123 Indeed, ifchargebacks rates are unusually high, a bank might refuse to provide service tothe merchant entirely, effectively denying the merchant access to the Visa andMasterCard system.124 To the extent that merchants do contest claims, the clearimpression in the industry is that the dispute resolution system is weightedheavily in favor of the consumer.125 For example, one executive explained:

[B]ecause this is a consumer-driven society and Visa and MasterCard areprimarily driven by the cardholder’s side—all the issuing banks are the oneswho sit on the board with Visa and MasterCard—I would say [it’s] just [the]law of averages[; I] would say that most of those would rule on behalf of thecardholder.126

119. See Butts Interview, supra note 69, at 1 (reporting an overall chargeback rate of 0.4% on alltransactions); Confrey Interview, supra note 71, at 8–9 (explaining that fraud claims and consumer-complaint claims each are about one half of all chargebacks); Klebe Interview, supra note 69, at 6(reporting an overall rate of 0.005%).

120. See Confrey Interview, supra note 71, at 9 (“[T]he merchant is usually better off in businesspractice. Going to the customer and trying to satisfy the customer rather than trying to deal through theacquirer and the issuer.”).

121. Klebe Interview, supra note 69, at 3.122. See Confrey Interview, supra note 71, at 9; Klebe Interview, supra note 69, at 3–4 (reporting a

fee that ranges from ten to fifteen dollars at the low end up to forty to fifty dollars on the high end).123. See Klebe Interview, supra note 69, at 3.124. See id.125. See Confrey Interview, supra note 71, at 10 (“The acquirers usually lose. Statistically, remem-

ber, that’s not a judgment as to whether their case is valid or not.”).126. Butts Interview, supra note 69 (transcript at 7). He reports that in the entire MasterCard system

there are only about ten such arbitration hearings each month. Butts Interview, supra note 69, at 7.


My information on that point is purely anecdotal and thus cannot resolve theempirical questions raised above. Moreover, it is somewhat ambiguous on thequestion of the merits of claims that consumers interpose.127 Furthermore,whatever the current level of frivolous claims might be, it is difficult to be surethat the level would not rise were TILA more broadly applicable.

Fortunately, a definitive analysis of that question is not necessary for therelatively limited thesis of this Essay. As explained above, my principal thesis isthat the existing framework draws an unjustified distinction between debit cardsand credit cards. Thus, the main point is that if the TILA provisions arejustifiable for credit cards, then they are also justifiable for debit cards. It wouldbe reasonable for the reader to conclude that the collapse of the debit/creditdistinction justifies a repeal of the TILA provisions in question without regardto the discussion in this section about the potential benefits of TILA.

It is important to point out some reasons why the TILA provisions may be anappropriate exercise of regulatory authority. The first point, and doubtless themost important, is that this is not an area in which consumer choice can beexpected to result in a set of perfectly efficient contract terms. As Cooter andRubin have explained in some detail, several related factors make market failurerelatively likely in the development of provisions in a contract with a financialinstitution for payment services. Among other things, the cost of negotiation islikely to exceed the potential benefits to the individual consumer; the consumeris unlikely to have complete information about the relevant differences betweenpotential card issuers; and even if the consumer does have such information, itis quite difficult for it to evaluate the tradeoff such differences should have bycomparison to the more salient features of the transactions: the fees that theissuer charges for its services.128

Furthermore, for the reasons explained above, a standard analysis frombehavioral economics justifies considerable skepticism about the likelihood thatthe consumer selecting a payment card from a menu of potential issuers willgive adequate weight in its selection to the issuer’s rules for dealing withcardholder-merchant disputes.129 The typical consumer is likely to focus muchmore closely on the interest rate and affinity benefits than it is on any informa-

127. That is not true not only because of the point noted above—that merchants give in toconsumers for reasons not directly related to the merits of the claims—but also because of the clearprevalence of a significant number of dubious fraud claims. See Butts Interview, supra note 69, at 4(describing “friendly fraud” claims as cases “that are claimed as fraudulent transactions but thecardholder really knows that they or a family member did in fact participate. They just because of theease of disputing the transaction, go ahead and dispute it.”). Those claims are particularly prevalent inthe area of “adult” information sold in the Internet. See id., at 8–9; Klebe Interview, supra note 69, at1–2.

128. See Cooter & Rubin, supra note 47, at 68–69 (discussing those difficulties and the limitedlikelihood that consumer choice will turn on the issuer’s policies for dealing with problem transac-tions); Norman I. Silber, Observing Reasonable Consumers: Cognitive Psychology, Consumer Behav-ior, and Consumer Law, 2 LOYOLA CONSUMER L. REP. 69, 72–73 (1990) (discussing cognitive problemsin consumer transactions).

129. See supra Part I.B.


tion it might be able to glean about the issuing bank’s practices in resolvingcardholder-merchant disputes.130

A natural regulatory response to that analysis is that a more effective disclo-sure regime could produce sufficient information to ensure that consumers makean informed choice between credit and debit cards. That could be done in any ofseveral ways: through consumer-awareness advertising by government agen-cies, by warnings included in monthly bank statements or even on the debit carditself, or perhaps by information at the point of sale. It is of course not possibleto be sure that such a scheme could not work, but any proponent of such a planmust recognize the general difficulty of using disclosure of information toovercome behavioral biases.131 This is particularly true here, where the informa-tion is relatively complex (the distinction between the different kinds of cardsand the legal import of that distinction) and when the people most in need of theinformation are not highly educated and sophisticated but rather ordinaryconsumers.132 Proponents of disclosure in this context need to recognize thatregulatory systems that might be effective when sophisticated intermediariescan use the information to drive market-clearing prices, as in the securitiesindustry, are not likely to be effective in altering the atomistic behavior ofindividual consumers. The best example from existing experience probably isthe home mortgage market, in which the federal government has waged aconcerted effort to disclose relevant information to consumers, but by allinformed accounts has had little positive effect.133 Thus, although I cannot besure that a disclosure regime could not foster informed choice, the situationdoes justify skepticism.

More broadly, giving consumers more detailed information about the incoher-ent distinctions that characterize the existing system does not address theconcerns discussed above. To be sure, a perfect and costless disclosure systemaddressed to perfectly rational consumers might result in perfect choices withrespect to reversibility, but in any real-world scenario there will be considerablecosts and imperfections. Thus, even if we can preserve a menu that allowsconsumers and merchants to select from different reversibility rules, the systemwould be better than it is now if that menu was coupled with default rules thatcame closer to reaching the optimal results than the rules currently in place

130. See William N. Eskridge, Jr., One Hundred Years of Ineptitude: The Need for Mortgage RulesConsistent with the Economic and Psychological Dynamics of the Home Sale and Loan Transaction, 70VA. L. REV. 1083, 1117–18 (1984) (making a similar point about home-mortgage loan transactions);Bar-Gill, supra note 90, at 1395–1411 (discussing the relation between cognitive problems and thestructure of credit card transactions); Cooter & Rubin, supra note 47, at 122 (discussing cognitiveproblems in transactions between consumers and financial institutions).

131. For a general discussion, see William M. Sage, Regulating Through Information: DisclosureLaws and American Health Care, 99 COLUM. L. REV. 1701, 1729–30 (1999).

132. See Sage, supra note 131, at 1728–29 (making a similar point about health-care disclosures).133. See Eskridge, supra note 130, at 1128–54.


do.134 That ideal menu, in turn, would be supplemented with policies designedto ensure that consumers understood the relevant rules, so that they could selectthe appropriate payment system for their situation. A coherent system shouldboth lower the costs of disclosure—because it would be easier to explain—andlower the costs of selection among systems—because the default rules would becloser to optimal.

Another important concern is the differing classes of consumers that will holddebit and credit cards. Although credit cards have now penetrated quite thor-oughly into our society, a significant portion of the population, 24% of familiesby one recent estimate, do not have a credit card even now.135 That 24%,however, is disproportionately composed of poor and minority families. Thus,for example, only 35% of families with an income below $10,000 have a creditcard, but 98% of families with an income over $100,000 have one.136 Consum-ers that do not have a credit card do not choose a payment system without thereversibility option because they are imprudent; they choose it because they arepoor. That is not a problem to be solved by enhancements in the informationavailable to cardholders.

The information effects of the rule also are relevant. As suggested above, thefinancial system currently relies on chargeback information in part to identifymerchants that engage in unfair practices and consequently to exclude themfrom the credit-card system. Thus, the information generated by a consumer’schargeback decision itself provides a positive externality to the system. That istrue whether the individual chargeback claim has merit or not because charge-back claims as a whole are likely to correlate inversely with merchant quality.That positive externality should have some weight in analysis of the overalleffect of the system.

Finally, even the most optimistic believer in the effectiveness of reputationalsanctions on merchant actors137 must acknowledge the existence of a large classof consumer retail purchases in which reputational sanctions will give themerchant little or no incentive to provide fair redress to the consumer. Mostobviously, the merchant might be insolvent or close to insolvency and thus haveno concern for future transactions.138 More generally, merchants remote fromthe cardholder’s residence—engaged in sales to tourists and the like—rationallymight doubt the possibility that a disappointed customer would be able to

134. This point is analogous to the common problem of “sticky” default rules. See, e.g., Ronald J.Mann, Contracts—Only with Consent, 152 U. PA. L. REV. 1873 (2004).

135. See Demos, Borrowing to Make Ends Meet: The Growth of Credit Card Debt in the ‘90s, at 10(2003), available at http://www.demos-usa.org/pubs/borrowing_to_make_ends_meet.pdf (last visitedMay 19, 2004).

136. See Demos, supra note 135, at 10.137. As it happens, I am quite optimistic about the effectiveness of reputational sanctions as a

general matter. See Mann, supra note 64, at 2252–57 (discussing factors that make reputationalsanctioning effective).

138. Recall the transaction from which this article began, in which the retailer failed withoutperforming the obligations for which my colleague had paid.


impose any significant reputational harm on the merchant, however intransigentthe merchant might be.139

IV. APPLICATIONS

A. FACE-TO-FACE TRANSACTIONS

Collectively, the preceding points are enough to persuade me in the absenceof any countervailing empirical evidence that the TILA protections reflect anappropriate intervention to redress a significant imbalance in litigation capacity.The question, then, is what policies that conclusion implies. Albeit with sometrepidation, I propose an extension of those protections to all face-to-face retailtransactions that use cards that access either a deposit account or a line ofcredit.140 My proposal would differ from existing law both in extending theexisting TILA rule to debit cards, and also in removing the limitation on thatrule to local transactions that currently appears in TILA Section 170(b).141

One obvious concern is that the extension of the reversibility rule to the debitcard context will increase the costs of debit cards to those that use them. If so,the reform might alter the relative desirability of the products in significantways. From one perspective, that is not a reason for concern. The only reasonthat it might alter the costs significantly is if there is a significant volume ofchargeback activity, which suggests that there are a significant number oftransactions in which consumers currently lack effective recourse. On the otherhand, as discussed above, there is the empirical possibility that a significantlevel of chargebacks might reflect abusive consumer conduct rather than dishon-est merchant conduct. For the reasons discussed above, however, I think itunlikely that there will be a sufficiently large volume of chargebacks to affectpricing significantly, largely because the volume of payment-reversing charge-backs in the credit-card system now is quite small.142

It would be easy to push that rationale farther to transactions that use cash orchecks, but for several reasons that seems to me unwise. First, the mostdetermined consumer advocate could not credibly claim that the TILA policiesshould be extended to all transactions and all payment instruments. There is

139. That line of reasoning, of course, suggests the perversity of the provision in TILA Section170(b), which excludes such transactions from the consumer’s right to withhold payment under TILASection 170(a). See supra notes 71–72 and accompanying text (discussing that exclusion); see alsoMANN, supra note 1, at 120 (Problem 6.6) (presenting that point).

140. The purpose of that somewhat convoluted formulation is to leave open the possibility that someof the newly developing card systems such as prepaid cards and gift cards properly might be treated asproviding a cash-equivalent payment option. Because the systems for issuing and using those cards aredeveloping so rapidly, I leave the details of any such exception for another day. My main point,however, is that I do not intend to exclude the possibility of a cash-equivalent electronic paymentoption from face-to-face transactions. I simply think that the debit card as it has developed should notbe viewed as such a thing.

141. See supra note 71 (discussing that rule).142. See supra note 119 (discussing current chargeback rates in credit card system).


some category of transactions for which reversibility would be completelyincompatible with the expectations of the parties, if only because the purchasedobject is one for which the likelihood of a dispute is too low to justify such aremedy. I think, for example, of the New York City street vendor from whom Imight buy a newspaper or beverage.

To accommodate those transactions, it seems crucial that the system providea menu of payment options with different reversibility attributes. To be sure,one could argue that the law should redress problems of leverage imbalancethrough the imposition of a single set of rules on all consumer payment systems.That seems, however, to go too far. For one thing, as long as there is a menu ofpayment systems that includes reasonably practical options with reversibilityand without reversibility, the system imposes relatively small costs on mer-chants and customers because merchants can insist on irrevocable payment, ifthey wish, simply by insisting on cash or cash equivalents. Consumers con-cerned about reversibility, conversely, can shift patronage to merchants thataccept reversible payment systems or choose to use those systems when mer-chants accept both types.

The discussion above, however, convinces me that the need for a menu ofoptions justifies a distinction between credit and debit cards primarily becausethat distinction is not likely to present a realistic option in the current milieu.There certainly is a tradeoff between a system in which consumers chooseamong the various payment systems based on their reversibility attributes and asystem in which legal rules establish a single reversibility rule that governs alltransactions of a particular type without regard to the payment mechanism thatthe consumer selects. The first system reflects the difficulty any governmentwould face in defining the proper universe of transactions to be covered by anyparticular rule of reversibility and relies on market forces to produce theappropriate coverage. The second system reflects the likelihood that marketforces, even when supported by government disclosure and information efforts,cannot produce an appropriate result. The discussion above suggests to me that,in this context at least, it is plausible to follow the second route—having the govern-ment articulate mandatory rules of reversibility based on the type of transaction.

The next question is how to define the boundaries of the covered transaction.The logic of the framework of this essay is that distinctions among reversibilityattributes should depend at least in part on the nature of the transaction. Thus, inan ideal world, perhaps the legal rule that described transactions in whichpayment was reversible would define the relevant class by reference to variousfeatures of the transaction such as the type of goods and the purchase price.

Notwithstanding the existing provisions in TILA Section 170(b), it does notmake sense to exclude remote transactions.143 The premise of that exclusion inexisting law is that credit card issuers will have arrangements with local

143. By “remote transactions” I mean transactions that are conducted more than 100 miles from thecardholder’s residence and outside the cardholder’s state.


merchants that will make it easy for them to verify the quality of thosemerchants and to recover from those merchants in the event of a dispute.144 Inmodern commerce, however, that argument has no empirical support. For onething, the modern chargeback system allows the issuer to recover the fundspromptly without regard to the merchant’s location.145 If anything, the imbal-ance in litigation capabilities discussed in the previous section is likely to beexacerbated in remote transactions where consumers are likely to be incapableof effective recourse. That is not to say that consumers would not abuse such arule when they deal with remote merchants. It is simply to recognize that thereare dishonest merchants as well as dishonest consumers; that reputationalconstraints are unlikely to constrain either of them; and that as a class the honestmerchants will be better situated to deal with the dishonest consumers than thehonest consumers will be situated to deal with the dishonest merchants. As withmuch of this discussion, empirical information about the nature of chargebackclaims would influence my views, but it is difficult to see how that informationcould be generated in a reliable and cost effective way.

At the same time, I also think there is much to be said for a small-dollarexclusion from the reversibility right, with a floor in the range of $50.146 Theidea is that there are certain sales transactions in which the optimal rulecertainly is to deny reversibility—the transaction to purchase a newspaper orpiece of fruit from a street vendor providing an obvious benchmark. In thatcontext, it might be that claims to reverse payment would more commonly beabusive than in the context of more significant purchases, if only because it isdifficult to imagine a legitimate basis for reversing payment in that context.147

At the same time, the relative burdens of costs and benefits are different in thatcontext because the administrative costs of reversing payment and dealing withthe chargeback are likely to be relatively fixed, while the benefits to theconsumer—the loss that the consumer otherwise would have borne if it could

144. See Brandel & Leonard, supra note 68, at 1064–68. Brandel & Leonard (who wrote recom-mended rules that ultimately became the statute now codified in TILA) justified the basic reversibilityrule by analogy to negotiability doctrine, which bars holder-in-due-course status for lenders that areclosely connected with merchants. See id., at 1042–44. It would follow naturally from that justificationthat issuers should be free from merchant-based claims of their cardholders when the merchants areremote from the issuer. As the text suggests, however, I do not think the analogy to negotiability shoulddrive the rules at this late date. See generally Mann, supra note 35 (discussing the declining importanceof negotiability doctrine).

145. See supra notes 69–70 and accompanying text.146. Brandel & Leonard argued for a $50 floor in their pre-TILA work advocating reversibility,

Brandel & Leonard, supra note 68, at 1059–64, but it was not included in the statute as enacted. I usethat number for discussion here because it is the figure that TILA uses as the floor for claims ofunauthorized transactions, but I recognize that it is a much smaller exclusion than the $50 exclusion in1971 dollars that Brandel & Leonard contemplated.

147. Remember, I am not discussing claims that transactions are unauthorized; see supra textaccompanying notes 17–18.


not reverse the payment—are diminished.148 Thus, I am inclined to support a$50 floor for the reversibility right, parallel to the $50 floor in existing law forclaims that transactions are not authorized.

One seeming inconsistency in my analysis is that I would define the universeof transactions also by reference to the particular payment system, including themajor existing card-based systems but excluding cash and checks. I rest thataspect of the analysis on practical concerns, motivated in part by the desire topreserve some menu of reversibility options. At least in the United States,149

where non-cash payment systems have so penetrated the economy, a ruleproviding substantially lower protections for the consumer in a cash transactionis unlikely to lead to a shift toward cash-only merchants. For the reasonssummarized above, street merchants and casual restaurants aside, merchants inour economy rarely find it profitable to insist on cash as a payment medium.150

Also, and perhaps more importantly, it is difficult to imagine a practical systemfor involuntarily retracting cash payments: cash transactions obviously lack thenetwork of relationships that leaves a payment path over which the consumercan retract the payment. To restate what should be obvious by now, thatproblem does not apply to debit cards because of the increasing parallelism ofthe systems for clearing debit and credit card transactions.

Checks are a harder problem, and thus a closer call. On the one hand,consumers that use checks to make purchases often use them in transactions of asize and type quite similar to the transactions in which they use credit cards,with the choice depending more on personal predilection than anything else.151

Thus, there is every reason to expect that consumers that purchase with checkswould benefit as much from a right to retract payment as consumers thatpurchase with credit cards.

On the other hand, a rule permitting check purchasers to retract paymentswould be almost as complicated as a rule that benefited cash purchasers becauseit would require substantial revisions to Article 4 of the Uniform CommercialCode and Regulation CC.152 Perhaps more importantly, checks resemble cash inthe sense that the path of a cleared check is not nearly so easily retraced as thepath of a completed credit card transaction. The issuing bank in a credit card

148. If the burdens of such a rule exceed its benefits, then a statute imposing a $50 floor onreversibility claims might make it easier for low-price merchants to accept payment cards. To the extentthat the rule lowered their transaction costs, it would provide a net social benefit.

149. As I write this passage in a Tokyo hotel—in a country where cash payment seems to be muchmore common, for much larger transactions, than it is in the United States—I am conscious of thecultural contingency of the analysis that follows.

150. More importantly, those merchants that do refuse credit cards rarely accept debit cards, so analteration of the legal rules related to debit cards would be unlikely to affect their transactions.

151. In 2002, the average consumer retail purchase with a check was $77, with a credit card $64,with a debit card $36, and with cash $22. See Consumer Payment Systems, supra note 45, at 6.

152. Regulation CC, 12 C.F.R. Part 229 (2004), sets out regulations promulgated under the Expe-dited Funds Availability Act that regulate the check-collection process. See generally MANN, supra note1, at 60–66 (discussing briefly the provisions of Regulation CC relevant to the check-collectionprocess).


transaction has an all but immediate, almost cost-free, and usually certainmechanism to retract funds when it unwinds a credit card transaction. Incontrast, whatever rights the law might grant, the bank on which a check isdrawn often would have significant doubts about its ability to retract the fundsthrough the system. Most obviously, it could recover the money only if thedepositor’s account contained sufficient funds to cover the retracted check.Because merchants generally leave the credit card system only upon insolvency,that is not likely to be a problem of similar significance in unwinding credit cardtransactions.

Moreover, even if it were practical to extend retraction rights into thechecking system, it is not clear that such a rule would produce benefits toconsumers that plausibly could justify the complexity of the necessary reforms.For one thing, the steadily diminishing level of retail check usage suggests thatimprovement of the system in this respect hardly justifies the complicatingfeatures that would be required. That is particularly true in light of the readyavailability of the card-based payment system as an option for the “savvy”purchaser to whom this issue is important. Credit and debit cards have suffi-ciently penetrated the marketplace that almost any consumer can choose to usea card-based payment system.153 Again, my instinct is that, at least in ourcountry, merchants would not respond to the increase of consumer protectionsfor debit cards by refusing to take credit and debit cards.154 If those protectionswere a concern, the merchants already would be refusing to take credit cards.More specifically, if TILA protections were a significant concern for merchants,they would not be objecting to Visa and MasterCard policies that require themto accept debit cards.155

In sum, I tentatively propose a mandatory rule of reversibility for both debitand credit cards. Checks and cash would be excluded, but transactions that useddebit cards would include rights to interpose defenses to payment for a period oftime in the range of 60 days, analogous to the period practically available undercurrent TILA-based credit card practices.156 The rule would apply to local and

153. Although almost a quarter of families do not have credit cards, see supra notes 135–136 andaccompanying text, many of those families have debit cards, particularly among families who havechecks. The rise of prepaid cards, which are available over the counter at local convenience stores, andthus do not require a checking account, may make the cardless consumer even rarer. See ConsumerPayment Systems, supra note 45, at 6 (reporting that prepaid cards were used for 1.99 billiontransactions in the United States in 2002 (1.65% of all retail payments), with an average amount of$27).

154. It might be that it would cause fewer consumers to use checks. Given the relatively high costsof check-processing, however, that effect in itself might provide a significant benefit to society.

155. See Bennett, supra note 103 (discussing such a complaint interposed by, among others,Wal-Mart and Sears); Bennett, Wal-Mart’s Bull Eyes Master Card, CREDIT CARD MANAGEMENT, Jan.2004, at 6 (reporting decision of Wal-Mart to refuse MasterCard’s PIN-less debit product, although itwill continue to accept MasterCard’s credit card product).

156. I recognize that academic calls for payment systems reform based on large-framed structuralsimilarities have not always been well received, as evidenced by the dismal reception of the UniformNew Payments Code. See, e.g., Gillette, supra note 12, at 219–20 (discussing the failure of the Uniform


nonlocal transactions, but not to small transactions below an arbitrary thresholdin the vicinity of $50.

B. INTERNET TRANSACTIONS

The most prominent change in payments over the next decade surely will bethe continuing shift of the market from face-to-face retail transactions toInternet transactions. From a market of truly insignificant proportions five yearsago to a trillion-dollar market in the next few years, the growth is nothing shortof exponential.157

Almost buried in the bustle of news stories about e-commerce is the questionof how consumers will pay for their e-purchases. Presently, those transactionsare paid for almost exclusively with credit cards.158 Thus, absent some changeof infrastructure, those transactions might be thought to raise no questions ofinterest for the subject of this Essay.

That conclusion, however, would be too facile. As it happens, the credit cardin its current form is not well suited for the broad run of transactions that areexpected to populate the Internet in the years to come. Three concerns aresalient. One is the privacy issue: credit card transactions result in the collectionof a considerable body of valuable information by the issuing bank. Manyobservers worry that a shift of purchasing to the Internet will lead to a shiftaway from cash purchases toward credit card purchases and to a consequentincrease in the ability of banks to collect and, more to the point, disseminateinformation about consumer purchasing power.159

Second, because of its relatively high per-unit charges, the credit card transac-tion is not well suited to small “micro-transactions.” Many observers believethat a significant market in the electronic commerce of the future will consist ofthe purchase and sale of individual pieces of information ranging from weatherreports, to sports information, to music.160 If those transactions are priced on aper-unit basis and paid for contemporaneously, the credit card would notprovide a workable payment device.161

Third, the credit card as used on the Internet seems to be particularly

New Payments Code and its implication for reform proposals in the area). I hope that my effort to relyon contextual detail and functional consistency distinguishes my work from that project, but recognizethe possibility that it does not. I note that the rule on which I focus in this Essay is one that the UniformNew Payments Code left entirely to private contract. See supra note 51.

157. See, e.g., Keith Regan, Holiday E-Tail Sales Set Records Despite Performance Woes, E-Commerce Times, Dec. 30, 2003, at http://www.ecommercetimes.com/perl/story/32491.html (reportingInternet retail sales of $16 billion during the 2003 holiday season from Nov. 1 to Dec. 19).

158. See Mann, supra note 44.159. See RONALD J. MANN & JANE KAUFMAN WINN, ELECTRONIC COMMERCE 584–85 (2d ed. 2005)

(discussing that problem).160. See Klebe Interview, supra note 69, at 17 (discussing music transactions that occur already).161. See MANN & WINN, supra note 159, at 585–86 (discussing that problem); see Klebe Interview,

supra note 69, at 15 (“Visa and MasterCard are actually geared towards transactions that are in like the$65–$80 range. They are not even really geared toward transactions down in the $10–$50 range. Butthey handle it reasonably well. Anything below that they just can’t do.”).


susceptible to fraud, which increases the cost of those transactions signifi-cantly.162 The difficulty is that those transactions, like mail-order transactions,offer the merchant no credible mechanism for verifying the identity of thepurported cardholder. Recognizing that problem, the card issuer in those transac-tions, unlike face-to-face transactions in which the card is “swiped,” leaves therisk of unauthorized transactions on the merchant.163

Because of the potential size of the Internet market, it should come as nosurprise that considerable effort is being devoted to a response to those prob-lems. The simplest response would be deployment of technology permitting thegeneral acceptance of debit cards on the Internet.164 More ambitious are theprojects attempting to develop a new electronic currency—generally referred toas electronic money. Using technology closely related to stored-value cards—and, ideally, compatible with them165—electronic money would allow consum-ers to pay merchants by transferring to them packages of electronic impulses(ecoins) that reflect value previously deposited with a reputable financial institu-tion. Because the ecoins would not be traceable to a particular consumer, thebank would not be able to develop records of the consumer’s transactions.Similarly, because the system would be entirely electronic, its per-transactioncharges would be sufficiently low to accommodate microtransactions.166 Lesstechnologically ambitious but more practical responses depend on aggregationof large numbers of transactions into a single billing event.167

The credit card industry, of course, could overcome those problems as well. Itcould upgrade its processing system to lower its transaction charges so as tomake credit cards practical for smaller transactions. Issuers also could refrainfrom collecting and disseminating purchasing information about their customersto third parties. At least to date, however, I am aware of nothing that suggeststhe industry is moving in either of those directions. Furthermore, and perhapsmost seriously in the long run, it is less clear what the industry could do toprevent the use of stolen credit card numbers in Internet transactions. Absentsome definitive mechanism for tying the individual transmitting information to

162. See MANN & WINN, supra note 159, at 581–84 (discussing concerns about fraud in Internettransactions and possible responses).

163. See id.; Confrey Interview, supra note 71, at 6–7 (discussing the historical development of thatdistinction in the credit card industry); Klebe Interview, supra note 69, at 8–9 (same).

164. See, e.g., Peter Lucas, How the Rest of the World Does Web Payments, CREDIT CARD MANAGE-MENT, Apr. 2003, at 24.

165. Although it seems an obvious advantage, it appears that the system being tested in Japan is oneof only a few systems yet put into operation that offers both stored-value and Internet capabilities for anelectronic payment device. See Yasushi Nakayama et al., An Electronic Money Scheme: A Proposal fora New Electronic Money Scheme Which Is Both Secure and Convenient 7–11 (Institute for Monetaryand Economic Studies, Bank of Japan, Discussion Paper No. 97-E-4) (describing that system).

166. For a general description of the mechanics of electronic-money systems, see MANN & WINN,supra note 159, at 614–20. For a more technical description, see David Chaum, Achieving ElectronicPrivacy, SCI. AM., Aug. 1992, at 96.

167. MANN & WINN, supra note 159, at 460–61.


the merchant to a particular card account,168 that problem seems likely toremain.169 Thus, the likelihood of a large-scale electronic-money system seemsgreater here than in the face-to-face context discussed above.

Although the stated purposes for developing such a system strike me asentirely salutary, one obvious side effect under the existing legal regime wouldbe to remove a large class of payment transactions from the protections grantedby TILA to current Internet transactions. But that result is just as much awindfall to the payment operator here as it is to the debit card issuer discussedin the preceding sections. Accordingly, it might be appropriate to extend thecharge-back rights described above to any Internet electronic payment systemas well, even if that system diverges from current practice sufficiently toabandon the use of a card-based account.

There are no obvious technological difficulties in that approach. Although apayor-anonymous system might not provide the bank with records of theparticular transactions of its customer, the customer that wishes to charge back atransaction need only forward to the bank the customer’s record of the transac-tion from the customer’s own computer. That record should include enoughinformation to allow the bank to match the transaction to the transaction itcleared from the merchant. At that point, the electronic-money system wouldresemble the credit card system, with the issuer in possession of informationthat would enable it to retract funds previously forwarded in payment of thetransaction. In sum, I doubt that it would be technologically difficult to includea charge-back option in an electronic-money system.170

The biggest problem with that approach is that it would make it quite difficultin the electronic milieu for a merchant to insist on a practical payment optionthat is irreversible. As discussed above, I see considerable value in permittingmerchants some flexibility in the type of payments that they are willing toaccept. In my view, however, that problem is less serious in the Internet context.For one thing, it is clear that the absence of such an option is not a majorproblem at present. Internet commerce grows rapidly now with no signs ofhindrance from the chargeback rights currently available to Internet purchas-

168. The most obvious solution to that problem would be a biometric device that would ensure that aperson present at the computer terminal had some characteristic that matched information residing onthe card. Even with the rise of biometric passports as a response to the September 11 attacks, such asolution remains some years away. See, e.g., Smart Travel, ECONOMIST, Mar. 12, 2005.

169. The credit card industry has initiated programs that would have consumers use PINs with creditcards on the Internet. See, e.g., https://usa.visa.com/personal/secure_with_visa/verified_by_visa.html(last visited May 18, 2004) (discussing the “Verified by Visa” program); http://www.mastercardmerchant.com/securecode/ (last visited May 18, 2004) (discussing MasterCard’s “SecureCode” program). If thoseprograms succeed, it is likely that fraud rates would fall significantly.

170. Informal conversations with officials at the Bank of Japan responsible for their system convinceme that this would be possible. The problem could be solved more directly, of course, by abandoningthe payor-anonymous features described above (as the Japanese system does), in which case theclearing banks could simply forward complete transaction records to the issuing bank when theyprocess payments. See Nakayama et al., supra note 165, at 10–11 (describing that process for handlingunauthorized transactions).


ers.171

I also am influenced here by the typical nature of Internet transactions, whichat least at the present time are likely to involve relatively large purchasesshipped remotely, a context in which the problem of leverage imbalance islikely to be particularly serious. A rule that does not, as a practical matter,permit irreversible payments is not in practice any different from the rulediscussed above which permits merchants to insist on irreversible paymentsonly if the merchants are willing to insist on cash for large purchases.

Finally, if a context appears in which a payment option of irreversibilitywould be important—micropurchases of information, for example—the exis-tence of the transactions will afford an opportunity for designing a rule thatwould accommodate such an option. It would be easy at that time, for example,for the appropriate regulatory authority, presumably the Federal Reserve, toadopt a transaction-bounded rule permitting merchants to make cash-like irrevers-ible payments in appropriate circumstances.172

It is difficult to fashion a payments policy for transactions that are onlybeginning to occur.173 Thus, my recommendations with respect to electronicpayments in this section must be taken much more tentatively than my analo-gous recommendations in the previous part. Still, the discussion itself has valueif only to pose the basic choice of whether payments in electronic transactionsshould be analyzed under rules that depend solely on the type of technology orunder rules that also depend on the type of transaction. I hope my discussionillustrates the virtue of the latter course.

V. CONCLUSION

I have argued that payments policy should be reexamined in a fundamental way.Once we see the source of those policies in the nature of the underlying transactions, itbecomes clear that the rules should be determined not only by the nature of thepayment systems that are used, but also by the nature of the transactions. Manymight disagree with my explication of that framework in the later parts of thisEssay, but I hope that most will agree that the framework itself can do much toilluminate the issues that should concern us in articulating payments policy.

171. That point can be overstated. There is, for example, a persistent problem with relatively highlevels of chargebacks in the adult-entertainment area, as customers disavow transactions in which theyin fact engaged. See Lauri Giesen, The Chargeback Squeeze, CREDIT CARD MANAGEMENT, Nov. 2003, at26 (general discussion of that problem). In my view, however, that problem is unique to an industry thatsells things customers do not wish to acknowledge buying. The problems of that industry should notdrive rules for more conventional commerce.

172. Such a rule, for example, might bar chargebacks for purchases of contemporaneously deliveredinformation with a cost per item below some threshold selected by regulation (such as a dollar ortwenty-five cents). I think such a rule would be appropriate even if there were not a general floor forreversibility claims. If there were such a floor, there would be no need for a special rule formicropurchases.

173. See Jane Kaufman Winn, Clash of the Titans: Regulating the Competition Between Establishedand Emerging Payment Systems, 14 BERKELEY TECH. L.J. 675, 708–09 (1999).


Making Sense of Payments Policy in the InformationAge

Documents

Transcript of Making Sense of Payments Policy in the InformationAge